{ "cells": [ { "attachments": {}, "cell_type": "markdown", "metadata": { "colab_type": "text", "id": "view-in-github" }, "source": [ "\"Open" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "zDG3hKiVmHUW" }, "source": [ "# How to use pgmuvi - a brief introduction\n", "\n" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "hEyug9B9LHZC" }, "source": [ "`pgmuvi` is a package to infer the multiwavelength behaviour of astronomical lightcurves using Gaussian Processes. It uses the `torch` and `gpytorch` libraries to define GP models, and the probabilistic programming language `pyro` to perform optimisation and sampling of the parameter space.\n", "\n", "## Why pgmuvi\n", "\n", "Many packages exist to interpret lightcurves with GPs, so why do we need another one? `pgmuvi` builds on the latest highly-optimised libraries which allow inference with arbitrary GP kernels. This makes it feasible to implement models that are very general, which is important in astronomy where lightcurves can have a wide range of different properties. However, this normally results in a speed penalty, since naive GPs scale with $\\mathcal{O}\\left(n^3\\right)$; to combat this, the libraries implement scalable approaches to GP inference, such as GPU computing. This allows both a wide range of GP kernels and fast inference, meaning that `pgmuvi` can be applied to arbitrary astronomical lightcurves without taking _that much_ longer than codes which are restricted to special cases but blisteringly fast in those cases.\n", "\n", "## What does pgmuvi do\n", "\n", "The objective is to understand what the PSD of the variability is, and in the case of (quasi-)periodic variables determine their dominant periods. To do this, `pgmuvi` uses a `SpectralMixtureKernel`, which models the PSD of the covariances of the GP as a Gaussian Mixture model. This is a very flexible model which is able to represent a wide range of different kinds of behaviour, so can be useful for a wide range of astronomical lightcurves.\n", "\n", "## How to use pgmuvi\n", "\n", "The rest of this notebook is basically a quickstart guide to `pgmuvi`. It generates sythetic observations, but you can replace these with reading in your own data and passing it to `pgmuvi`'s `Lightcurve` classes to apply it to your own data." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "# We will start by fixing the seed for reproducibility\n", "seed = 0\n", "import torch\n", "\n", "torch.manual_seed(seed)\n", "import numpy as np\n", "\n", "np.random.seed(seed)\n", "import random\n", "\n", "random.seed(seed)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Q1SU6ncv7BtF", "outputId": "196f7bc8-0df8-4ec0-95eb-bfb4356b9aa7" }, "outputs": [], "source": [ "try: # This won't work right now - instead clone the repository and `pip install -e .`\n", " import pgmuvi\n", "except (ImportError, ModuleNotFoundError):\n", " %pip install git+https://github.com/ICSM/pgmuvi.git\n", " import pgmuvi" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "G9zGgK85MpGK" }, "source": [ "Now we have imported/installed pgmuvi, we can use it to fit some data. For the purposes of this tutorial, we will generate two synthetic datasets - a single-wavelength lightcurve with very simple behaviour, and a second multi-wavelength one with some more complex patterns in it. It is important to note that the data must be in the form of `torch` tensors, rather than numpy arrays or other data types." ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "l7iXofH7Nbz_", "outputId": "c837d540-5ab7-49e8-d439-ab031c2bea84" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "True period: 178.17964606037768 days\n", "Simulating for 8.006325564606936 periods\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "c:\\Users\\peter\\anaconda3\\envs\\pgmuvi_dev\\Lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", " from .autonotebook import tqdm as notebook_tqdm\n" ] } ], "source": [ "import gpytorch\n", "\n", "\"\"\" Let's generate some synthetic data from a perturbed sine curve\n", " but on the same time sampling as the real data\"\"\"\n", "\n", "P = np.random.uniform(30, 300) # 137. #Days!\n", "print(\"True period: \", P, \" days\")\n", "n_data = 400\n", "jd_min = 2450000\n", "n_periods = np.random.uniform(3, 10)\n", "jd_max = jd_min + P * (n_periods)\n", "print(\"Simulating for \", n_periods, \" periods\")\n", "\n", "# train_mag =\n", "# train_mag = train_mag + 0.1*torch.randn_like(train_mag)\n", "# train_mag_err = 0.1*train_mag\n", "\n", "period_guess = P * (\n", " np.random.uniform() + 0.5\n", ") # 147 #this number is in the same units as our original input.\n", "\n", "# generate data from a simple case - superimpose two sine curves and add noise\n", "timestamps_1d = torch.Tensor(\n", " np.random.uniform(jd_min, jd_max, size=n_data)\n", ") # generate random x data here\n", "fluxes_1d = torch.sin(timestamps_1d * (2 * np.pi / P)) # generate random y data here\n", "fluxes_1d += 0.1 * torch.randn_like(fluxes_1d)\n", "flux_err_1d = 0.1 * fluxes_1d.abs()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "jP7ZjHrTOaWp" }, "source": [ "Now that we've got some simple data, we can feed it to pgmuvi and see how it handles it." ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "id": "L1GzKnLJOzHh" }, "outputs": [], "source": [ "from pgmuvi.lightcurve import Lightcurve\n", "\n", "lightcurve_1d = Lightcurve(\n", " timestamps_1d, fluxes_1d, yerr=flux_err_1d, xtransform=\"minmax\"\n", ")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "m06s-TKiPtfL" }, "source": [ "The `xtransform` argument is important. The computing time and numerical stability of this GP model depends somewhat on the range of frequencies it is asked to compute the PSD on, so it is common to transform the inputs to some simple interval that ensures the frequencies are easier. Astronomical timeseries often cover very long times, and therefore very low frequencies if left in their original units (e.g. Julian Dates), which are particularly difficult to compute on. By default, `Lightcurve` does not apply any transform. However, it is often useful to transform the inputs to a simpler interval to improve numerical stability and computing time. For example, astronomical timeseries often cover very long times (e.g. Julian Dates), leading to very low frequencies if left in their original units, which are particularly difficult to compute on. The user can specify a transform such as ``minmax`` (which rescales the data to the closed interval $\\left[0, 1\\right]$), ``zscore``, or ``robust_score`` by passing the appropriate string to the `xtransform` argument.\n", "\n", "Now that we have a lightcurve, we can try to fit it:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "qV5QuLYdPwTx", "outputId": "18f0c945-f9b2-4951-a3b7-d75e1d883072" }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "c:\\Users\\peter\\anaconda3\\envs\\pgmuvi_dev\\Lib\\site-packages\\gpytorch\\likelihoods\\noise_models.py:148: NumericalWarning: Very small noise values detected. This will likely lead to numerical instabilities. Rounding small noise values up to 0.0001.\n", " warnings.warn(\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "likelihood.second_noise_covar.noise: tensor([0.3674])\n", "mean_module.constant: 0.02001810073852539\n", "covar_module.mixture_weights: tensor([0.7348])\n", "covar_module.mixture_means: tensor([[[1.4098]]])\n", "covar_module.mixture_scales: tensor([[[0.0016]]])\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 3000/3000 [00:41<00:00, 72.86it/s]\n" ] } ], "source": [ "fit = lightcurve_1d.fit(\n", " model=\"1D\", likelihood=\"learn\", num_mixtures=1, training_iter=3000, miniter=3000\n", ")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "lwIguoa3QAD2" }, "source": [ "this is a very simple case, and we have not specified any of the extra keywords that `fit` exposes. You will probably want to specify an initial guess of the GP hyperparameters, and the default number of components in the mixture model for the PSD is 4. Each mixture has a mean, scale and weight for the Gaussian, and there is a mean function (a constant mean) and an extra free parameter to learn additional noise in the input data.\n", "\n", "Now that the fit is completed, you can manually inspect the results" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "OwkXyfaIQGLi", "outputId": "fa0ece25-3f8a-4a1b-dcc9-bb9ded5bd96d" }, "outputs": [ { "data": { "text/plain": [ "{'loss': [array(1.153896, dtype=float32),\n", " array(1.143073, dtype=float32),\n", " array(1.1406492, dtype=float32),\n", " array(1.1368848, dtype=float32),\n", " array(1.1255214, dtype=float32),\n", " array(1.1199832, dtype=float32),\n", " array(1.1201514, dtype=float32),\n", " array(1.1178579, dtype=float32),\n", " array(1.1130016, dtype=float32),\n", " array(1.1138129, dtype=float32),\n", " array(1.1141769, dtype=float32),\n", " array(1.1113707, dtype=float32),\n", " array(1.108514, dtype=float32),\n", " array(1.1102849, dtype=float32),\n", " array(1.1075708, dtype=float32),\n", " array(1.1077479, dtype=float32),\n", " array(1.1069926, dtype=float32),\n", " array(1.1058103, dtype=float32),\n", " array(1.1101999, dtype=float32),\n", " array(1.112484, dtype=float32),\n", " array(1.1067532, dtype=float32),\n", " array(1.0965791, dtype=float32),\n", " array(1.0925707, dtype=float32),\n", " array(1.103569, dtype=float32),\n", " array(1.1065767, dtype=float32),\n", " array(1.105345, dtype=float32),\n", " array(1.1006522, dtype=float32),\n", " array(1.1025436, dtype=float32),\n", " array(1.1044317, dtype=float32),\n", " array(1.1049088, dtype=float32),\n", " array(1.1074133, dtype=float32),\n", " array(1.0996994, dtype=float32),\n", " array(1.0963757, dtype=float32),\n", " array(1.1008737, dtype=float32),\n", " array(1.0998044, dtype=float32),\n", " array(1.0981877, dtype=float32),\n", " array(1.0942851, dtype=float32),\n", " array(1.0956832, dtype=float32),\n", " array(1.0951811, dtype=float32),\n", " array(1.0991848, dtype=float32),\n", " array(1.0920466, dtype=float32),\n", " array(1.0945289, dtype=float32),\n", " array(1.0940801, dtype=float32),\n", " array(1.0982529, dtype=float32),\n", " array(1.0987898, dtype=float32),\n", " array(1.0976567, dtype=float32),\n", " array(1.0967342, dtype=float32),\n", " array(1.0951467, dtype=float32),\n", " array(1.0967952, dtype=float32),\n", " array(1.0967152, dtype=float32),\n", " array(1.0985293, dtype=float32),\n", " array(1.0916181, dtype=float32),\n", " array(1.0932366, dtype=float32),\n", " array(1.0875055, dtype=float32),\n", " array(1.093496, dtype=float32),\n", " array(1.0967156, dtype=float32),\n", " array(1.0953624, dtype=float32),\n", " array(1.0955914, dtype=float32),\n", " array(1.0939391, dtype=float32),\n", " array(1.09375, dtype=float32),\n", " array(1.0965811, dtype=float32),\n", " array(1.0930268, dtype=float32),\n", " array(1.0947678, dtype=float32),\n", " array(1.0943193, dtype=float32),\n", " array(1.091792, dtype=float32),\n", " array(1.0915091, dtype=float32),\n", " array(1.0930612, dtype=float32),\n", " array(1.092877, dtype=float32),\n", " array(1.090107, dtype=float32),\n", " array(1.0911307, dtype=float32),\n", " array(1.0922725, dtype=float32),\n", " array(1.094522, dtype=float32),\n", " array(1.093755, dtype=float32),\n", " array(1.0922288, dtype=float32),\n", " array(1.0921553, dtype=float32),\n", " array(1.0925826, dtype=float32),\n", " array(1.0937881, dtype=float32),\n", " array(1.0886344, dtype=float32),\n", " array(1.0894684, dtype=float32),\n", " array(1.0934752, dtype=float32),\n", " array(1.0907232, dtype=float32),\n", " array(1.0942295, dtype=float32),\n", " array(1.0937228, dtype=float32),\n", " array(1.0939782, dtype=float32),\n", " array(1.0919222, dtype=float32),\n", " array(1.0909829, dtype=float32),\n", " array(1.0937004, dtype=float32),\n", " array(1.0922874, dtype=float32),\n", " array(1.0935522, dtype=float32),\n", " array(1.0880946, dtype=float32),\n", " array(1.0931975, dtype=float32),\n", " array(1.091478, dtype=float32),\n", " array(1.0938803, dtype=float32),\n", " array(1.0903724, dtype=float32),\n", " array(1.0921551, dtype=float32),\n", " array(1.088387, dtype=float32),\n", " array(1.0888425, dtype=float32),\n", " array(1.0909398, dtype=float32),\n", " array(1.0911161, dtype=float32),\n", " array(1.0924875, dtype=float32),\n", " array(1.09153, dtype=float32),\n", " array(1.0921338, dtype=float32),\n", " array(1.0928376, dtype=float32),\n", " array(1.0931891, dtype=float32),\n", " array(1.0887877, dtype=float32),\n", " array(1.0900208, dtype=float32),\n", " array(1.0923905, dtype=float32),\n", " array(1.0925597, dtype=float32),\n", " array(1.0905861, dtype=float32),\n", " array(1.0931207, dtype=float32),\n", " array(1.0930078, dtype=float32),\n", " array(1.0914532, dtype=float32),\n", " array(1.0921952, dtype=float32),\n", " array(1.0906551, dtype=float32),\n", " array(1.0920138, dtype=float32),\n", " array(1.092434, dtype=float32),\n", " array(1.0925673, dtype=float32),\n", " array(1.0920916, dtype=float32),\n", " array(1.0927856, dtype=float32),\n", " array(1.0921665, dtype=float32),\n", " array(1.0915232, dtype=float32),\n", " array(1.0911778, dtype=float32),\n", " array(1.0907081, dtype=float32),\n", " array(1.0925099, dtype=float32),\n", " array(1.0920393, dtype=float32),\n", " array(1.0923239, dtype=float32),\n", " array(1.0922754, dtype=float32),\n", " array(1.0921764, dtype=float32),\n", " array(1.0918211, dtype=float32),\n", " array(1.0910262, dtype=float32),\n", " array(1.0904001, dtype=float32),\n", " array(1.0894742, dtype=float32),\n", " array(1.0920057, dtype=float32),\n", " array(1.0897146, dtype=float32),\n", " array(1.0918385, dtype=float32),\n", " array(1.0904868, dtype=float32),\n", " array(1.092215, dtype=float32),\n", " array(1.0921112, dtype=float32),\n", " array(1.0880765, dtype=float32),\n", " array(1.0910511, dtype=float32),\n", " array(1.0911629, dtype=float32),\n", " array(1.0905867, dtype=float32),\n", " array(1.0917983, dtype=float32),\n", " array(1.0915079, dtype=float32),\n", " array(1.090254, dtype=float32),\n", " array(1.0906485, dtype=float32),\n", " array(1.090643, dtype=float32),\n", " array(1.0892649, dtype=float32),\n", " array(1.0915107, dtype=float32),\n", " array(1.0920697, dtype=float32),\n", " array(1.0888476, dtype=float32),\n", " array(1.0908675, dtype=float32),\n", " array(1.0911962, dtype=float32),\n", " array(1.0912068, dtype=float32),\n", " array(1.0909171, dtype=float32),\n", " array(1.090203, dtype=float32),\n", " array(1.0905243, dtype=float32),\n", " array(1.091708, dtype=float32),\n", " array(1.091327, dtype=float32),\n", " array(1.091328, dtype=float32),\n", " array(1.0906291, dtype=float32),\n", " array(1.0913434, dtype=float32),\n", " array(1.0895177, dtype=float32),\n", " array(1.0914273, dtype=float32),\n", " array(1.09166, dtype=float32),\n", " array(1.0917137, dtype=float32),\n", " array(1.0886441, dtype=float32),\n", " array(1.0909245, dtype=float32),\n", " array(1.0903449, dtype=float32),\n", " array(1.0902916, dtype=float32),\n", " array(1.0909266, dtype=float32),\n", " array(1.0910832, dtype=float32),\n", " array(1.091009, dtype=float32),\n", " array(1.0908777, dtype=float32),\n", " array(1.0905479, dtype=float32),\n", " array(1.0896057, dtype=float32),\n", " array(1.0890635, dtype=float32),\n", " array(1.0912733, dtype=float32),\n", " array(1.0879766, dtype=float32),\n", " array(1.0907042, dtype=float32),\n", " array(1.0911267, dtype=float32),\n", " array(1.0908501, dtype=float32),\n", " array(1.0906806, dtype=float32),\n", " array(1.090696, dtype=float32),\n", " array(1.0901757, dtype=float32),\n", " array(1.0913033, dtype=float32),\n", " array(1.0904163, dtype=float32),\n", " array(1.0915395, dtype=float32),\n", " array(1.0900662, dtype=float32),\n", " array(1.090179, dtype=float32),\n", " array(1.089579, dtype=float32),\n", " array(1.0914984, dtype=float32),\n", " array(1.0912491, dtype=float32),\n", " array(1.0910093, dtype=float32),\n", " array(1.0905002, dtype=float32),\n", " array(1.0911126, dtype=float32),\n", " array(1.0904553, dtype=float32),\n", " array(1.0909791, dtype=float32),\n", " array(1.090763, dtype=float32),\n", " array(1.0912093, dtype=float32),\n", " array(1.0892038, dtype=float32),\n", " array(1.089961, dtype=float32),\n", " array(1.0901746, dtype=float32),\n", " array(1.0909951, dtype=float32),\n", " array(1.0908782, dtype=float32),\n", " array(1.0895706, dtype=float32),\n", " array(1.0912999, dtype=float32),\n", " array(1.0914626, dtype=float32),\n", " array(1.0915406, dtype=float32),\n", " array(1.091448, dtype=float32),\n", " array(1.0911913, dtype=float32),\n", " array(1.0903684, dtype=float32),\n", " array(1.0900038, dtype=float32),\n", " array(1.0913435, dtype=float32),\n", " array(1.0903639, dtype=float32),\n", " array(1.0896294, dtype=float32),\n", " array(1.0885303, dtype=float32),\n", " array(1.0906343, dtype=float32),\n", " array(1.0895904, dtype=float32),\n", " array(1.0907727, dtype=float32),\n", " array(1.0905238, dtype=float32),\n", " array(1.0890033, dtype=float32),\n", " array(1.0905652, dtype=float32),\n", " array(1.0904415, dtype=float32),\n", " array(1.0911558, dtype=float32),\n", " array(1.090877, dtype=float32),\n", " array(1.0904354, dtype=float32),\n", " array(1.0901878, dtype=float32),\n", " array(1.0904459, dtype=float32),\n", " array(1.0906428, dtype=float32),\n", " array(1.08895, dtype=float32),\n", " array(1.0890507, dtype=float32),\n", " array(1.0910563, dtype=float32),\n", " array(1.0913882, dtype=float32),\n", " array(1.0907147, dtype=float32),\n", " array(1.091105, dtype=float32),\n", " array(1.0910393, dtype=float32),\n", " array(1.0907797, dtype=float32),\n", " array(1.0882325, dtype=float32),\n", " array(1.0873122, dtype=float32),\n", " array(1.0894921, dtype=float32),\n", " array(1.0898407, dtype=float32),\n", " array(1.0909638, dtype=float32),\n", " array(1.0911306, dtype=float32),\n", " array(1.0914989, dtype=float32),\n", " array(1.0886711, dtype=float32),\n", " array(1.0902439, dtype=float32),\n", " array(1.0909153, dtype=float32),\n", " array(1.0911744, dtype=float32),\n", " array(1.0912396, dtype=float32),\n", " array(1.0906154, dtype=float32),\n", " array(1.0896711, dtype=float32),\n", " array(1.0902661, dtype=float32),\n", " array(1.0903945, dtype=float32),\n", " array(1.0909485, dtype=float32),\n", " array(1.0890723, dtype=float32),\n", " array(1.0882745, dtype=float32),\n", " array(1.0906913, dtype=float32),\n", " array(1.0905949, dtype=float32),\n", " array(1.090855, dtype=float32),\n", " array(1.0910275, dtype=float32),\n", " array(1.0897877, dtype=float32),\n", " array(1.090948, dtype=float32),\n", " array(1.091077, dtype=float32),\n", " array(1.0890115, dtype=float32),\n", " array(1.088988, dtype=float32),\n", " array(1.090099, dtype=float32),\n", " array(1.0889766, dtype=float32),\n", " array(1.0903612, dtype=float32),\n", " array(1.090884, dtype=float32),\n", " array(1.0910697, dtype=float32),\n", " array(1.0913708, dtype=float32),\n", " array(1.0905371, dtype=float32),\n", " array(1.0911528, dtype=float32),\n", " array(1.0901835, dtype=float32),\n", " array(1.09051, dtype=float32),\n", " array(1.0914197, dtype=float32),\n", " array(1.0907367, dtype=float32),\n", " array(1.0913651, dtype=float32),\n", " array(1.0904449, dtype=float32),\n", " array(1.0891055, dtype=float32),\n", " array(1.0909785, dtype=float32),\n", " array(1.0886502, dtype=float32),\n", " array(1.0896196, dtype=float32),\n", " array(1.090346, dtype=float32),\n", " array(1.0913564, dtype=float32),\n", " array(1.0915756, dtype=float32),\n", " array(1.0908796, dtype=float32),\n", " array(1.0902139, dtype=float32),\n", " array(1.0897262, dtype=float32),\n", " array(1.0899125, dtype=float32),\n", " array(1.090404, dtype=float32),\n", " array(1.0890998, dtype=float32),\n", " array(1.0880363, dtype=float32),\n", " array(1.0907627, dtype=float32),\n", " array(1.0910863, dtype=float32),\n", " array(1.0904343, dtype=float32),\n", " array(1.0907338, dtype=float32),\n", " array(1.0892744, dtype=float32),\n", " array(1.0896661, dtype=float32),\n", " array(1.0910782, dtype=float32),\n", " array(1.0901542, dtype=float32),\n", " array(1.0903662, dtype=float32),\n", " array(1.089597, dtype=float32),\n", " array(1.0900158, dtype=float32),\n", " array(1.0909233, dtype=float32),\n", " array(1.0912111, dtype=float32),\n", " array(1.0900885, dtype=float32),\n", " array(1.0894339, dtype=float32),\n", " array(1.0891049, dtype=float32),\n", " array(1.0894208, dtype=float32),\n", " array(1.0909567, dtype=float32),\n", " array(1.0901184, dtype=float32),\n", " array(1.0909804, dtype=float32),\n", " array(1.0909123, dtype=float32),\n", " array(1.0911897, dtype=float32),\n", " array(1.0907704, dtype=float32),\n", " array(1.0905471, dtype=float32),\n", " array(1.0907899, dtype=float32),\n", " array(1.0910615, dtype=float32),\n", " array(1.0908316, dtype=float32),\n", " array(1.0897342, dtype=float32),\n", " array(1.0909781, dtype=float32),\n", " array(1.0906285, dtype=float32),\n", " array(1.0905366, dtype=float32),\n", " array(1.090278, dtype=float32),\n", " array(1.0899583, dtype=float32),\n", " array(1.0907927, dtype=float32),\n", " array(1.0908011, dtype=float32),\n", " array(1.08979, dtype=float32),\n", " array(1.0904422, dtype=float32),\n", " array(1.0908481, dtype=float32),\n", " array(1.0911595, dtype=float32),\n", " array(1.0909636, dtype=float32),\n", " array(1.090642, dtype=float32),\n", " array(1.0903093, dtype=float32),\n", " array(1.0876262, dtype=float32),\n", " array(1.0903567, dtype=float32),\n", " array(1.0912066, dtype=float32),\n", " array(1.0911851, dtype=float32),\n", " array(1.0904025, dtype=float32),\n", " array(1.0909806, dtype=float32),\n", " array(1.0902072, dtype=float32),\n", " array(1.090816, dtype=float32),\n", " array(1.0897067, dtype=float32),\n", " array(1.0898792, dtype=float32),\n", " array(1.0895205, dtype=float32),\n", " array(1.0896502, dtype=float32),\n", " array(1.0906587, dtype=float32),\n", " array(1.0891097, dtype=float32),\n", " array(1.0902302, dtype=float32),\n", " array(1.0907922, dtype=float32),\n", " array(1.0911921, dtype=float32),\n", " array(1.0913011, dtype=float32),\n", " array(1.0909936, dtype=float32),\n", " array(1.0901074, dtype=float32),\n", " array(1.0905262, dtype=float32),\n", " array(1.0903008, dtype=float32),\n", " array(1.0901834, dtype=float32),\n", " array(1.0906844, dtype=float32),\n", " array(1.0890073, dtype=float32),\n", " array(1.0892731, dtype=float32),\n", " array(1.0910279, dtype=float32),\n", " array(1.0907204, dtype=float32),\n", " array(1.090623, dtype=float32),\n", " array(1.0905657, dtype=float32),\n", " array(1.0896529, dtype=float32),\n", " array(1.0900052, dtype=float32),\n", " array(1.0905708, dtype=float32),\n", " array(1.0906053, dtype=float32),\n", " array(1.0909572, dtype=float32),\n", " array(1.0912421, dtype=float32),\n", " array(1.0909069, dtype=float32),\n", " array(1.0907325, dtype=float32),\n", " array(1.091264, dtype=float32),\n", " array(1.091182, dtype=float32),\n", " array(1.0907458, dtype=float32),\n", " array(1.0902932, dtype=float32),\n", " array(1.090752, dtype=float32),\n", " array(1.0907663, dtype=float32),\n", " array(1.0905423, dtype=float32),\n", " array(1.091078, dtype=float32),\n", " array(1.0910945, dtype=float32),\n", " array(1.0905316, dtype=float32),\n", " array(1.0887545, dtype=float32),\n", " array(1.0893319, dtype=float32),\n", " array(1.0897943, dtype=float32),\n", " array(1.0904243, dtype=float32),\n", " array(1.0901763, dtype=float32),\n", " array(1.0896046, dtype=float32),\n", " array(1.0898212, dtype=float32),\n", " array(1.0891962, dtype=float32),\n", " array(1.0897734, dtype=float32),\n", " array(1.0906485, dtype=float32),\n", " array(1.0909265, dtype=float32),\n", " array(1.0899678, dtype=float32),\n", " array(1.0893197, dtype=float32),\n", " array(1.0900396, dtype=float32),\n", " array(1.0901668, dtype=float32),\n", " array(1.0908425, dtype=float32),\n", " array(1.0908039, dtype=float32),\n", " array(1.0906441, dtype=float32),\n", " array(1.0895224, dtype=float32),\n", " array(1.0900747, dtype=float32),\n", " array(1.0905962, dtype=float32),\n", " array(1.0907382, dtype=float32),\n", " array(1.089951, dtype=float32),\n", " array(1.0902766, dtype=float32),\n", " array(1.0901725, dtype=float32),\n", " array(1.0907167, dtype=float32),\n", " array(1.0905285, dtype=float32),\n", " array(1.0899762, dtype=float32),\n", " array(1.0901623, dtype=float32),\n", " array(1.0904686, dtype=float32),\n", " array(1.0899569, dtype=float32),\n", " array(1.0895946, dtype=float32),\n", " array(1.0896473, dtype=float32),\n", " array(1.0906, dtype=float32),\n", " array(1.0907633, dtype=float32),\n", " array(1.0906762, dtype=float32),\n", " array(1.0903618, dtype=float32),\n", " array(1.0905958, dtype=float32),\n", " array(1.0906063, dtype=float32),\n", " array(1.0900817, dtype=float32),\n", " array(1.0905626, dtype=float32),\n", " array(1.091137, dtype=float32),\n", " array(1.0905288, dtype=float32),\n", " array(1.0900908, dtype=float32),\n", " array(1.0895041, dtype=float32),\n", " array(1.0886775, dtype=float32),\n", " array(1.0897022, dtype=float32),\n", " array(1.0897956, dtype=float32),\n", " array(1.0879482, dtype=float32),\n", " array(1.0891616, dtype=float32),\n", " array(1.0899608, dtype=float32),\n", " array(1.0900426, dtype=float32),\n", " array(1.0908221, dtype=float32),\n", " array(1.0910915, dtype=float32),\n", " array(1.0913393, dtype=float32),\n", " array(1.0913444, dtype=float32),\n", " array(1.0911084, dtype=float32),\n", " array(1.0906959, dtype=float32),\n", " array(1.0898155, dtype=float32),\n", " array(1.0888605, dtype=float32),\n", " array(1.0883162, dtype=float32),\n", " array(1.0895194, dtype=float32),\n", " array(1.0903039, dtype=float32),\n", " array(1.0896523, dtype=float32),\n", " array(1.0904106, dtype=float32),\n", " array(1.0901393, dtype=float32),\n", " array(1.0899882, dtype=float32),\n", " array(1.0910523, dtype=float32),\n", " array(1.0910007, dtype=float32),\n", " array(1.0903033, dtype=float32),\n", " array(1.0908324, dtype=float32),\n", " array(1.0908651, dtype=float32),\n", " array(1.0902644, dtype=float32),\n", " array(1.0903261, dtype=float32),\n", " array(1.0903103, dtype=float32),\n", " array(1.0905294, dtype=float32),\n", " array(1.0903032, dtype=float32),\n", " array(1.0908518, dtype=float32),\n", " array(1.0912735, dtype=float32),\n", " array(1.0910544, dtype=float32),\n", " array(1.0908366, dtype=float32),\n", " array(1.0908673, dtype=float32),\n", " array(1.0902321, dtype=float32),\n", " array(1.0902226, dtype=float32),\n", " array(1.0907207, dtype=float32),\n", " array(1.0909708, dtype=float32),\n", " array(1.0910343, dtype=float32),\n", " array(1.0911248, dtype=float32),\n", " array(1.0909219, dtype=float32),\n", " array(1.0904862, dtype=float32),\n", " array(1.0899866, dtype=float32),\n", " array(1.0902727, dtype=float32),\n", " array(1.0909175, dtype=float32),\n", " array(1.0910172, dtype=float32),\n", " array(1.090796, dtype=float32),\n", " array(1.0905858, dtype=float32),\n", " array(1.0901726, dtype=float32),\n", " array(1.0900602, dtype=float32),\n", " array(1.0904319, dtype=float32),\n", " array(1.0902478, dtype=float32),\n", " array(1.0905129, dtype=float32),\n", " array(1.0910339, dtype=float32),\n", " array(1.090569, dtype=float32),\n", " array(1.0900755, dtype=float32),\n", " array(1.0902951, dtype=float32),\n", " array(1.08986, dtype=float32),\n", " array(1.0899642, dtype=float32),\n", " array(1.090222, dtype=float32),\n", " array(1.090185, dtype=float32),\n", " array(1.0902708, dtype=float32),\n", " array(1.0909243, dtype=float32),\n", " array(1.091186, dtype=float32),\n", " array(1.090739, dtype=float32),\n", " array(1.0902499, dtype=float32),\n", " array(1.0903019, dtype=float32),\n", " array(1.0901958, dtype=float32),\n", " array(1.0903488, dtype=float32),\n", " array(1.0905019, dtype=float32),\n", " array(1.0907134, dtype=float32),\n", " array(1.0906992, dtype=float32),\n", " array(1.0907437, dtype=float32),\n", " array(1.0908312, dtype=float32),\n", " array(1.0908282, dtype=float32),\n", " array(1.0908116, dtype=float32),\n", " array(1.090921, dtype=float32),\n", " array(1.0910147, dtype=float32),\n", " array(1.0908493, dtype=float32),\n", " array(1.0905575, dtype=float32),\n", " array(1.0906242, dtype=float32),\n", " array(1.0903976, dtype=float32),\n", " array(1.0886698, dtype=float32),\n", " array(1.0884346, dtype=float32),\n", " array(1.0901594, dtype=float32),\n", " array(1.090623, dtype=float32),\n", " array(1.0903821, dtype=float32),\n", " array(1.089931, dtype=float32),\n", " array(1.0901246, dtype=float32),\n", " array(1.0903682, dtype=float32),\n", " array(1.0904098, dtype=float32),\n", " array(1.090571, dtype=float32),\n", " array(1.0901091, dtype=float32),\n", " array(1.0896811, dtype=float32),\n", " array(1.0897508, dtype=float32),\n", " array(1.0900772, dtype=float32),\n", " array(1.0902759, dtype=float32),\n", " array(1.090621, dtype=float32),\n", " array(1.090757, dtype=float32),\n", " array(1.0908023, dtype=float32),\n", " array(1.0909088, dtype=float32),\n", " array(1.0906141, dtype=float32),\n", " array(1.0900638, dtype=float32),\n", " array(1.0899858, dtype=float32),\n", " array(1.0901791, dtype=float32),\n", " array(1.0902747, dtype=float32),\n", " array(1.0906595, dtype=float32),\n", " array(1.090833, dtype=float32),\n", " array(1.0905738, dtype=float32),\n", " array(1.0905038, dtype=float32),\n", " array(1.0907096, dtype=float32),\n", " array(1.0907866, dtype=float32),\n", " array(1.090715, dtype=float32),\n", " array(1.0907304, dtype=float32),\n", " array(1.0907979, dtype=float32),\n", " array(1.0907122, dtype=float32),\n", " array(1.0904818, dtype=float32),\n", " array(1.0901663, dtype=float32),\n", " array(1.0899864, dtype=float32),\n", " array(1.0901042, dtype=float32),\n", " array(1.090376, dtype=float32),\n", " array(1.0904686, dtype=float32),\n", " array(1.0904987, dtype=float32),\n", " array(1.0903976, dtype=float32),\n", " array(1.090439, dtype=float32),\n", " array(1.0905684, dtype=float32),\n", " array(1.0902461, dtype=float32),\n", " array(1.0899215, dtype=float32),\n", " array(1.0901772, dtype=float32),\n", " array(1.0904679, dtype=float32),\n", " array(1.0903809, dtype=float32),\n", " array(1.0899646, dtype=float32),\n", " array(1.0895363, dtype=float32),\n", " array(1.0898466, dtype=float32),\n", " array(1.0903244, dtype=float32),\n", " array(1.0904894, dtype=float32),\n", " array(1.0904061, dtype=float32),\n", " array(1.0898594, dtype=float32),\n", " array(1.0893441, dtype=float32),\n", " array(1.0894703, dtype=float32),\n", " array(1.0897803, dtype=float32),\n", " array(1.0899479, dtype=float32),\n", " array(1.0901079, dtype=float32),\n", " array(1.0903589, dtype=float32),\n", " array(1.0902209, dtype=float32),\n", " array(1.0898869, dtype=float32),\n", " array(1.0896881, dtype=float32),\n", " array(1.0896076, dtype=float32),\n", " array(1.0901084, dtype=float32),\n", " array(1.090731, dtype=float32),\n", " array(1.0909035, dtype=float32),\n", " array(1.0905553, dtype=float32),\n", " array(1.0894297, dtype=float32),\n", " array(1.0887747, dtype=float32),\n", " array(1.0897348, dtype=float32),\n", " array(1.0905592, dtype=float32),\n", " array(1.0904471, dtype=float32),\n", " array(1.0903188, dtype=float32),\n", " array(1.0905958, dtype=float32),\n", " array(1.0907489, dtype=float32),\n", " array(1.090558, dtype=float32),\n", " array(1.0903822, dtype=float32),\n", " array(1.0902793, dtype=float32),\n", " array(1.0902705, dtype=float32),\n", " array(1.0901613, dtype=float32),\n", " array(1.0896453, dtype=float32),\n", " array(1.0887183, dtype=float32),\n", " array(1.0888066, dtype=float32),\n", " array(1.0900716, dtype=float32),\n", " array(1.0905842, dtype=float32),\n", " array(1.0903965, dtype=float32),\n", " array(1.0903902, dtype=float32),\n", " array(1.0906355, dtype=float32),\n", " array(1.090915, dtype=float32),\n", " array(1.0909005, dtype=float32),\n", " array(1.0903903, dtype=float32),\n", " array(1.0897934, dtype=float32),\n", " array(1.0894575, dtype=float32),\n", " array(1.0894642, dtype=float32),\n", " array(1.0896621, dtype=float32),\n", " array(1.0898979, dtype=float32),\n", " array(1.0897328, dtype=float32),\n", " array(1.0896131, dtype=float32),\n", " array(1.0902904, dtype=float32),\n", " array(1.0908401, dtype=float32),\n", " array(1.0909603, dtype=float32),\n", " array(1.0909992, dtype=float32),\n", " array(1.0909145, dtype=float32),\n", " array(1.0905455, dtype=float32),\n", " array(1.08986, dtype=float32),\n", " array(1.0895227, dtype=float32),\n", " array(1.0899231, dtype=float32),\n", " array(1.0903631, dtype=float32),\n", " array(1.0905235, dtype=float32),\n", " array(1.0905303, dtype=float32),\n", " array(1.0903332, dtype=float32),\n", " array(1.0903423, dtype=float32),\n", " array(1.0906682, dtype=float32),\n", " array(1.0908742, dtype=float32),\n", " array(1.0907289, dtype=float32),\n", " array(1.0901114, dtype=float32),\n", " array(1.0895925, dtype=float32),\n", " array(1.0899829, dtype=float32),\n", " array(1.0904121, dtype=float32),\n", " array(1.0900306, dtype=float32),\n", " array(1.0898314, dtype=float32),\n", " array(1.0902752, dtype=float32),\n", " array(1.0908023, dtype=float32),\n", " array(1.0910834, dtype=float32),\n", " array(1.0910517, dtype=float32),\n", " array(1.0907689, dtype=float32),\n", " array(1.0903629, dtype=float32),\n", " array(1.0901799, dtype=float32),\n", " array(1.0904382, dtype=float32),\n", " array(1.0906136, dtype=float32),\n", " array(1.0901741, dtype=float32),\n", " array(1.0898602, dtype=float32),\n", " array(1.0903362, dtype=float32),\n", " array(1.0908418, dtype=float32),\n", " array(1.0908022, dtype=float32),\n", " array(1.0906483, dtype=float32),\n", " array(1.0906988, dtype=float32),\n", " array(1.0907912, dtype=float32),\n", " array(1.090816, dtype=float32),\n", " array(1.090908, dtype=float32),\n", " array(1.0909299, dtype=float32),\n", " array(1.0907384, dtype=float32),\n", " array(1.0905107, dtype=float32),\n", " array(1.0904354, dtype=float32),\n", " array(1.0905522, dtype=float32),\n", " array(1.0906861, dtype=float32),\n", " array(1.0905417, dtype=float32),\n", " array(1.0901904, dtype=float32),\n", " array(1.0899848, dtype=float32),\n", " array(1.0899566, dtype=float32),\n", " array(1.0900763, dtype=float32),\n", " array(1.090453, dtype=float32),\n", " array(1.0908537, dtype=float32),\n", " array(1.0909717, dtype=float32),\n", " array(1.0908189, dtype=float32),\n", " array(1.0906199, dtype=float32),\n", " array(1.0906278, dtype=float32),\n", " array(1.0907664, dtype=float32),\n", " array(1.0907016, dtype=float32),\n", " array(1.0902492, dtype=float32),\n", " array(1.0896769, dtype=float32),\n", " array(1.0895871, dtype=float32),\n", " array(1.0899502, dtype=float32),\n", " array(1.0902473, dtype=float32),\n", " array(1.0903478, dtype=float32),\n", " array(1.0904001, dtype=float32),\n", " array(1.0905595, dtype=float32),\n", " array(1.090736, dtype=float32),\n", " array(1.0907304, dtype=float32),\n", " array(1.0905288, dtype=float32),\n", " array(1.090185, dtype=float32),\n", " array(1.089962, dtype=float32),\n", " array(1.090078, dtype=float32),\n", " array(1.0903349, dtype=float32),\n", " array(1.0904847, dtype=float32),\n", " array(1.0903043, dtype=float32),\n", " array(1.0899097, dtype=float32),\n", " array(1.0899347, dtype=float32),\n", " array(1.0903696, dtype=float32),\n", " array(1.0905904, dtype=float32),\n", " array(1.0904346, dtype=float32),\n", " array(1.0902959, dtype=float32),\n", " array(1.0902829, dtype=float32),\n", " array(1.0901128, dtype=float32),\n", " array(1.089742, dtype=float32),\n", " array(1.0894494, dtype=float32),\n", " array(1.0896753, dtype=float32),\n", " array(1.0902569, dtype=float32),\n", " array(1.0906299, dtype=float32),\n", " array(1.0907534, dtype=float32),\n", " array(1.0907071, dtype=float32),\n", " array(1.0905142, dtype=float32),\n", " array(1.0903914, dtype=float32),\n", " array(1.0903788, dtype=float32),\n", " array(1.090368, dtype=float32),\n", " array(1.0904068, dtype=float32),\n", " array(1.0904591, dtype=float32),\n", " array(1.0904229, dtype=float32),\n", " array(1.0902082, dtype=float32),\n", " array(1.089808, dtype=float32),\n", " array(1.0895795, dtype=float32),\n", " array(1.0896237, dtype=float32),\n", " array(1.0897188, dtype=float32),\n", " array(1.0899079, dtype=float32),\n", " array(1.0900708, dtype=float32),\n", " array(1.0900246, dtype=float32),\n", " array(1.0900276, dtype=float32),\n", " array(1.0901756, dtype=float32),\n", " array(1.0902412, dtype=float32),\n", " array(1.0900927, dtype=float32),\n", " array(1.0899776, dtype=float32),\n", " array(1.0902001, dtype=float32),\n", " array(1.0905484, dtype=float32),\n", " array(1.0906409, dtype=float32),\n", " array(1.0904462, dtype=float32),\n", " array(1.0901742, dtype=float32),\n", " array(1.0901217, dtype=float32),\n", " array(1.090339, dtype=float32),\n", " array(1.0904922, dtype=float32),\n", " array(1.0904126, dtype=float32),\n", " array(1.0903543, dtype=float32),\n", " array(1.0904809, dtype=float32),\n", " array(1.0906699, dtype=float32),\n", " array(1.0907658, dtype=float32),\n", " array(1.0905517, dtype=float32),\n", " array(1.0898234, dtype=float32),\n", " array(1.0891182, dtype=float32),\n", " array(1.0892442, dtype=float32),\n", " array(1.090004, dtype=float32),\n", " array(1.0905391, dtype=float32),\n", " array(1.0905734, dtype=float32),\n", " array(1.0904373, dtype=float32),\n", " array(1.0904365, dtype=float32),\n", " array(1.0904704, dtype=float32),\n", " array(1.0902673, dtype=float32),\n", " array(1.0898561, dtype=float32),\n", " array(1.0896322, dtype=float32),\n", " array(1.0898006, dtype=float32),\n", " array(1.0901558, dtype=float32),\n", " array(1.0904673, dtype=float32),\n", " array(1.0906799, dtype=float32),\n", " array(1.0907254, dtype=float32),\n", " array(1.0906116, dtype=float32),\n", " array(1.0904231, dtype=float32),\n", " array(1.0902734, dtype=float32),\n", " array(1.0902916, dtype=float32),\n", " array(1.0904415, dtype=float32),\n", " array(1.0905567, dtype=float32),\n", " array(1.0904495, dtype=float32),\n", " array(1.0901335, dtype=float32),\n", " array(1.089992, dtype=float32),\n", " array(1.0902433, dtype=float32),\n", " array(1.0905882, dtype=float32),\n", " array(1.0907496, dtype=float32),\n", " array(1.0907463, dtype=float32),\n", " array(1.0906225, dtype=float32),\n", " array(1.0904864, dtype=float32),\n", " array(1.0904975, dtype=float32),\n", " array(1.090582, dtype=float32),\n", " array(1.0905573, dtype=float32),\n", " array(1.0904964, dtype=float32),\n", " array(1.0905766, dtype=float32),\n", " array(1.0907502, dtype=float32),\n", " array(1.0908635, dtype=float32),\n", " array(1.0908581, dtype=float32),\n", " array(1.0907454, dtype=float32),\n", " array(1.0905986, dtype=float32),\n", " array(1.0905207, dtype=float32),\n", " array(1.090554, dtype=float32),\n", " array(1.0906292, dtype=float32),\n", " array(1.0906609, dtype=float32),\n", " array(1.0906566, dtype=float32),\n", " array(1.0906526, dtype=float32),\n", " array(1.090531, dtype=float32),\n", " array(1.0901382, dtype=float32),\n", " array(1.0895996, dtype=float32),\n", " array(1.0892426, dtype=float32),\n", " array(1.0892032, dtype=float32),\n", " array(1.0894481, dtype=float32),\n", " array(1.0899013, dtype=float32),\n", " array(1.0903196, dtype=float32),\n", " array(1.0904373, dtype=float32),\n", " array(1.0903845, dtype=float32),\n", " array(1.0904183, dtype=float32),\n", " array(1.0905391, dtype=float32),\n", " array(1.0906444, dtype=float32),\n", " array(1.0906875, dtype=float32),\n", " array(1.0905615, dtype=float32),\n", " array(1.090193, dtype=float32),\n", " array(1.0898237, dtype=float32),\n", " array(1.089756, dtype=float32),\n", " array(1.0898479, dtype=float32),\n", " array(1.0897937, dtype=float32),\n", " array(1.0896144, dtype=float32),\n", " array(1.0894454, dtype=float32),\n", " array(1.0893904, dtype=float32),\n", " array(1.0895735, dtype=float32),\n", " array(1.0899246, dtype=float32),\n", " array(1.0901457, dtype=float32),\n", " array(1.0900464, dtype=float32),\n", " array(1.0897882, dtype=float32),\n", " array(1.0896397, dtype=float32),\n", " array(1.0896597, dtype=float32),\n", " array(1.0898095, dtype=float32),\n", " array(1.0900716, dtype=float32),\n", " array(1.0903462, dtype=float32),\n", " array(1.0904924, dtype=float32),\n", " array(1.0905155, dtype=float32),\n", " array(1.0904107, dtype=float32),\n", " array(1.0900905, dtype=float32),\n", " array(1.0896955, dtype=float32),\n", " array(1.0895389, dtype=float32),\n", " array(1.0896987, dtype=float32),\n", " array(1.0898283, dtype=float32),\n", " array(1.0896199, dtype=float32),\n", " array(1.0893501, dtype=float32),\n", " array(1.0894446, dtype=float32),\n", " array(1.0898206, dtype=float32),\n", " array(1.090063, dtype=float32),\n", " array(1.0900679, dtype=float32),\n", " array(1.0900687, dtype=float32),\n", " array(1.0902207, dtype=float32),\n", " array(1.0904949, dtype=float32),\n", " array(1.090785, dtype=float32),\n", " array(1.0909427, dtype=float32),\n", " array(1.0908508, dtype=float32),\n", " array(1.0906012, dtype=float32),\n", " array(1.0904284, dtype=float32),\n", " array(1.09046, dtype=float32),\n", " array(1.0905782, dtype=float32),\n", " array(1.0905259, dtype=float32),\n", " array(1.0901303, dtype=float32),\n", " array(1.0894918, dtype=float32),\n", " array(1.0890552, dtype=float32),\n", " array(1.089157, dtype=float32),\n", " array(1.089629, dtype=float32),\n", " array(1.0899568, dtype=float32),\n", " array(1.0898895, dtype=float32),\n", " array(1.0897193, dtype=float32),\n", " array(1.0898355, dtype=float32),\n", " array(1.0902646, dtype=float32),\n", " array(1.0906731, dtype=float32),\n", " array(1.0906929, dtype=float32),\n", " array(1.0902556, dtype=float32),\n", " array(1.0897069, dtype=float32),\n", " array(1.0894984, dtype=float32),\n", " array(1.0897549, dtype=float32),\n", " array(1.0902039, dtype=float32),\n", " array(1.0904738, dtype=float32),\n", " array(1.0904677, dtype=float32),\n", " array(1.0902297, dtype=float32),\n", " array(1.0898962, dtype=float32),\n", " array(1.0897353, dtype=float32),\n", " array(1.0898824, dtype=float32),\n", " array(1.0902238, dtype=float32),\n", " array(1.0905436, dtype=float32),\n", " array(1.0907735, dtype=float32),\n", " array(1.0909681, dtype=float32),\n", " array(1.0910661, dtype=float32),\n", " array(1.0909283, dtype=float32),\n", " array(1.0905304, dtype=float32),\n", " array(1.090048, dtype=float32),\n", " array(1.089767, dtype=float32),\n", " array(1.0898204, dtype=float32),\n", " array(1.0900477, dtype=float32),\n", " array(1.0901488, dtype=float32),\n", " array(1.0900317, dtype=float32),\n", " array(1.0898538, dtype=float32),\n", " array(1.0897517, dtype=float32),\n", " array(1.089713, dtype=float32),\n", " array(1.0896196, dtype=float32),\n", " array(1.0893695, dtype=float32),\n", " array(1.0889813, dtype=float32),\n", " array(1.0886343, dtype=float32),\n", " array(1.0885844, dtype=float32),\n", " array(1.0889186, dtype=float32),\n", " array(1.0894502, dtype=float32),\n", " array(1.0899594, dtype=float32),\n", " array(1.0903639, dtype=float32),\n", " array(1.090717, dtype=float32),\n", " array(1.0910512, dtype=float32),\n", " array(1.0913246, dtype=float32),\n", " array(1.0914923, dtype=float32),\n", " array(1.0915433, dtype=float32),\n", " array(1.0915006, dtype=float32),\n", " array(1.0913395, dtype=float32),\n", " array(1.0910895, dtype=float32),\n", " array(1.090851, dtype=float32),\n", " array(1.0906705, dtype=float32),\n", " array(1.0905244, dtype=float32),\n", " array(1.0903546, dtype=float32),\n", " array(1.0901735, dtype=float32),\n", " array(1.0900491, dtype=float32),\n", " array(1.0900235, dtype=float32),\n", " array(1.090055, dtype=float32),\n", " array(1.0900894, dtype=float32),\n", " array(1.0901434, dtype=float32),\n", " array(1.0902379, dtype=float32),\n", " array(1.090369, dtype=float32),\n", " array(1.090493, dtype=float32),\n", " array(1.0905774, dtype=float32),\n", " array(1.0906405, dtype=float32),\n", " array(1.0906863, dtype=float32),\n", " array(1.0907044, dtype=float32),\n", " array(1.0907032, dtype=float32),\n", " array(1.0907129, dtype=float32),\n", " array(1.0907401, dtype=float32),\n", " array(1.0907482, dtype=float32),\n", " array(1.090709, dtype=float32),\n", " array(1.0906377, dtype=float32),\n", " array(1.0905777, dtype=float32),\n", " array(1.0905623, dtype=float32),\n", " array(1.0905929, dtype=float32),\n", " array(1.0906376, dtype=float32),\n", " array(1.090642, dtype=float32),\n", " array(1.0905781, dtype=float32),\n", " array(1.0904788, dtype=float32),\n", " array(1.0904056, dtype=float32),\n", " array(1.0903927, dtype=float32),\n", " array(1.0904309, dtype=float32),\n", " array(1.0904875, dtype=float32),\n", " array(1.0905428, dtype=float32),\n", " array(1.0905887, dtype=float32),\n", " array(1.0906001, dtype=float32),\n", " array(1.0905461, dtype=float32),\n", " array(1.0903823, dtype=float32),\n", " array(1.0900497, dtype=float32),\n", " array(1.0895635, dtype=float32),\n", " array(1.0890777, dtype=float32),\n", " array(1.088803, dtype=float32),\n", " array(1.0888618, dtype=float32),\n", " array(1.0892125, dtype=float32),\n", " array(1.0896935, dtype=float32),\n", " array(1.0901555, dtype=float32),\n", " array(1.0905213, dtype=float32),\n", " array(1.090751, dtype=float32),\n", " array(1.0908225, dtype=float32),\n", " array(1.0907552, dtype=float32),\n", " array(1.0906109, dtype=float32),\n", " array(1.0904592, dtype=float32),\n", " array(1.0903465, dtype=float32),\n", " array(1.0902786, dtype=float32),\n", " array(1.0902379, dtype=float32),\n", " array(1.0902245, dtype=float32),\n", " array(1.090252, dtype=float32),\n", " array(1.0903019, dtype=float32),\n", " array(1.0903102, dtype=float32),\n", " array(1.0902134, dtype=float32),\n", " array(1.090026, dtype=float32),\n", " array(1.0898523, dtype=float32),\n", " array(1.0897954, dtype=float32),\n", " array(1.0898792, dtype=float32),\n", " array(1.090057, dtype=float32),\n", " array(1.0902667, dtype=float32),\n", " array(1.0904659, dtype=float32),\n", " array(1.0906274, dtype=float32),\n", " array(1.0907226, dtype=float32),\n", " array(1.09072, dtype=float32),\n", " array(1.0906076, dtype=float32),\n", " array(1.0904025, dtype=float32),\n", " array(1.0901613, dtype=float32),\n", " array(1.0899752, dtype=float32),\n", " array(1.089928, dtype=float32),\n", " array(1.0900404, dtype=float32),\n", " array(1.0902482, dtype=float32),\n", " array(1.0904332, dtype=float32),\n", " array(1.0904938, dtype=float32),\n", " array(1.0904062, dtype=float32),\n", " array(1.0902151, dtype=float32),\n", " array(1.0899911, dtype=float32),\n", " array(1.0898147, dtype=float32),\n", " array(1.0897474, dtype=float32),\n", " array(1.0897951, dtype=float32),\n", " array(1.0899026, dtype=float32),\n", " array(1.0899796, dtype=float32),\n", " array(1.0899377, dtype=float32),\n", " array(1.0897455, dtype=float32),\n", " array(1.0894619, dtype=float32),\n", " array(1.0892104, dtype=float32),\n", " array(1.0891222, dtype=float32),\n", " array(1.0892591, dtype=float32),\n", " array(1.0895659, dtype=float32),\n", " array(1.0899104, dtype=float32),\n", " ...],\n", " 'delta_loss': [-0.010823011,\n", " -0.0024237633,\n", " -0.003764391,\n", " -0.011363387,\n", " -0.005538225,\n", " 0.00016820431,\n", " -0.0022934675,\n", " -0.004856348,\n", " 0.0008113384,\n", " 0.00036394596,\n", " -0.0028061867,\n", " -0.0028567314,\n", " 0.0017709732,\n", " -0.002714157,\n", " 0.000177145,\n", " -0.00075531006,\n", " -0.0011823177,\n", " 0.0043896437,\n", " 0.00228405,\n", " -0.005730748,\n", " -0.010174155,\n", " -0.0040084124,\n", " 0.010998368,\n", " 0.0030076504,\n", " -0.0012316704,\n", " -0.004692793,\n", " 0.0018913746,\n", " 0.0018881559,\n", " 0.00047707558,\n", " 0.002504468,\n", " -0.007713914,\n", " -0.0033236742,\n", " 0.004498005,\n", " -0.0010693073,\n", " -0.0016167164,\n", " -0.0039025545,\n", " 0.0013980865,\n", " -0.0005021095,\n", " 0.004003644,\n", " -0.007138133,\n", " 0.002482295,\n", " -0.00044882298,\n", " 0.004172802,\n", " 0.00053691864,\n", " -0.0011330843,\n", " -0.0009225607,\n", " -0.0015875101,\n", " 0.0016485453,\n", " -7.998943e-05,\n", " 0.001814127,\n", " -0.006911278,\n", " 0.0016185045,\n", " -0.005731106,\n", " 0.005990505,\n", " 0.0032196045,\n", " -0.0013531446,\n", " 0.00022900105,\n", " -0.00165236,\n", " -0.00018906593,\n", " 0.0028311014,\n", " -0.0035543442,\n", " 0.0017410517,\n", " -0.00044846535,\n", " -0.0025273561,\n", " -0.00028288364,\n", " 0.001552105,\n", " -0.00018417835,\n", " -0.0027700663,\n", " 0.0010237694,\n", " 0.0011417866,\n", " 0.0022494793,\n", " -0.00076699257,\n", " -0.0015262365,\n", " -7.343292e-05,\n", " 0.0004272461,\n", " 0.0012055635,\n", " -0.005153775,\n", " 0.0008339882,\n", " 0.0040068626,\n", " -0.0027520657,\n", " 0.0035063028,\n", " -0.0005066395,\n", " 0.0002553463,\n", " -0.0020560026,\n", " -0.00093925,\n", " 0.002717495,\n", " -0.0014129877,\n", " 0.0012648106,\n", " -0.0054576397,\n", " 0.005102873,\n", " -0.0017194748,\n", " 0.0024023056,\n", " -0.0035078526,\n", " 0.0017826557,\n", " -0.0037680864,\n", " 0.0004554987,\n", " 0.002097249,\n", " 0.00017631054,\n", " 0.0013713837,\n", " -0.000957489,\n", " 0.00060379505,\n", " 0.00070381165,\n", " 0.0003515482,\n", " -0.0044014454,\n", " 0.0012331009,\n", " 0.0023697615,\n", " 0.00016915798,\n", " -0.001973629,\n", " 0.002534628,\n", " -0.0001128912,\n", " -0.0015546083,\n", " 0.0007419586,\n", " -0.0015400648,\n", " 0.0013587475,\n", " 0.00042021275,\n", " 0.00013327599,\n", " -0.00047576427,\n", " 0.0006940365,\n", " -0.00061905384,\n", " -0.00064337254,\n", " -0.0003453493,\n", " -0.0004696846,\n", " 0.0018017292,\n", " -0.00047051907,\n", " 0.00028455257,\n", " -4.851818e-05,\n", " -9.894371e-05,\n", " -0.0003553629,\n", " -0.00079488754,\n", " -0.0006260872,\n", " -0.00092589855,\n", " 0.0025315285,\n", " -0.0022910833,\n", " 0.0021238327,\n", " -0.0013517141,\n", " 0.0017281771,\n", " -0.00010371208,\n", " -0.0040347576,\n", " 0.0029746294,\n", " 0.00011181831,\n", " -0.0005762577,\n", " 0.0012116432,\n", " -0.00029039383,\n", " -0.0012539625,\n", " 0.00039458275,\n", " -5.4836273e-06,\n", " -0.0013781786,\n", " 0.0022457838,\n", " 0.00055909157,\n", " -0.003222108,\n", " 0.0020198822,\n", " 0.00032866,\n", " 1.0609627e-05,\n", " -0.00028967857,\n", " -0.00071406364,\n", " 0.00032126904,\n", " 0.001183629,\n", " -0.0003809929,\n", " 1.0728836e-06,\n", " -0.00069892406,\n", " 0.00071430206,\n", " -0.0018256903,\n", " 0.0019096136,\n", " 0.00023269653,\n", " 5.364418e-05,\n", " -0.00306952,\n", " 0.0022803545,\n", " -0.00057959557,\n", " -5.3286552e-05,\n", " 0.0006350279,\n", " 0.0001565218,\n", " -7.414818e-05,\n", " -0.00013136864,\n", " -0.0003297329,\n", " -0.0009422302,\n", " -0.00054216385,\n", " 0.0022097826,\n", " -0.003296733,\n", " 0.0027276278,\n", " 0.00042247772,\n", " -0.00027656555,\n", " -0.00016951561,\n", " 1.5377998e-05,\n", " -0.00052022934,\n", " 0.0011276007,\n", " -0.0008870363,\n", " 0.0011231899,\n", " -0.0014733076,\n", " 0.00011277199,\n", " -0.00059998035,\n", " 0.0019193888,\n", " -0.00024926662,\n", " -0.00023984909,\n", " -0.00050902367,\n", " 0.0006123781,\n", " -0.00065732,\n", " 0.0005238056,\n", " -0.00021612644,\n", " 0.00044631958,\n", " -0.0020054579,\n", " 0.0007572174,\n", " 0.00021350384,\n", " 0.00082051754,\n", " -0.000116825104,\n", " -0.0013076067,\n", " 0.00172925,\n", " 0.00016272068,\n", " 7.7962875e-05,\n", " -9.262562e-05,\n", " -0.0002566576,\n", " -0.0008229017,\n", " -0.000364542,\n", " 0.001339674,\n", " -0.0009796619,\n", " -0.00073444843,\n", " -0.0010991096,\n", " 0.002104044,\n", " -0.0010439157,\n", " 0.0011823177,\n", " -0.000248909,\n", " -0.0015205145,\n", " 0.0015618801,\n", " -0.00012373924,\n", " 0.00071430206,\n", " -0.00027871132,\n", " -0.00044167042,\n", " -0.0002475977,\n", " 0.0002580881,\n", " 0.00019693375,\n", " -0.0016927719,\n", " 0.00010061264,\n", " 0.0020056963,\n", " 0.00033187866,\n", " -0.0006735325,\n", " 0.0003902912,\n", " -6.568432e-05,\n", " -0.00025963783,\n", " -0.002547145,\n", " -0.0009202957,\n", " 0.002179861,\n", " 0.00034856796,\n", " 0.0011231899,\n", " 0.0001667738,\n", " 0.0003682375,\n", " -0.0028277636,\n", " 0.0015728474,\n", " 0.0006713867,\n", " 0.0002590418,\n", " 6.520748e-05,\n", " -0.00062417984,\n", " -0.0009442568,\n", " 0.00059497356,\n", " 0.0001283884,\n", " 0.00055396557,\n", " -0.0018761158,\n", " -0.0007978678,\n", " 0.0024168491,\n", " -9.6440315e-05,\n", " 0.00026011467,\n", " 0.00017249584,\n", " -0.0012397766,\n", " 0.001160264,\n", " 0.00012898445,\n", " -0.0020654202,\n", " -2.360344e-05,\n", " 0.0011110306,\n", " -0.0011223555,\n", " 0.0013846159,\n", " 0.00052273273,\n", " 0.00018572807,\n", " 0.00030112267,\n", " -0.0008337498,\n", " 0.000615716,\n", " -0.00096929073,\n", " 0.00032651424,\n", " 0.0009096861,\n", " -0.00068295,\n", " 0.00062835217,\n", " -0.0009201765,\n", " -0.0013394356,\n", " 0.0018730164,\n", " -0.0023282766,\n", " 0.00096940994,\n", " 0.0007263422,\n", " 0.0010104179,\n", " 0.00021922588,\n", " -0.00069606304,\n", " -0.0006656647,\n", " -0.0004876852,\n", " 0.00018632412,\n", " 0.0004914999,\n", " -0.0013042688,\n", " -0.0010634661,\n", " 0.0027264357,\n", " 0.000323534,\n", " -0.0006519556,\n", " 0.00029945374,\n", " -0.0014593601,\n", " 0.00039172173,\n", " 0.001412034,\n", " -0.0009239912,\n", " 0.00021207333,\n", " -0.00076925755,\n", " 0.00041878223,\n", " 0.0009075403,\n", " 0.00028777122,\n", " -0.0011225939,\n", " -0.0006545782,\n", " -0.00032901764,\n", " 0.00031590462,\n", " 0.0015358925,\n", " -0.0008382797,\n", " 0.0008620024,\n", " -6.8068504e-05,\n", " 0.00027740002,\n", " -0.00041937828,\n", " -0.000223279,\n", " 0.00024282932,\n", " 0.00027155876,\n", " -0.00022983551,\n", " -0.0010974407,\n", " 0.0012439489,\n", " -0.00034964085,\n", " -9.191036e-05,\n", " -0.00025856495,\n", " -0.00031971931,\n", " 0.0008343458,\n", " 8.46386e-06,\n", " -0.0010111332,\n", " 0.000652194,\n", " 0.00040590763,\n", " 0.00031137466,\n", " -0.00019586086,\n", " -0.00032162666,\n", " -0.00033271313,\n", " -0.0026830435,\n", " 0.0027304888,\n", " 0.000849843,\n", " -2.1457672e-05,\n", " -0.000782609,\n", " 0.00057816505,\n", " -0.00077342987,\n", " 0.00060880184,\n", " -0.0011093616,\n", " 0.00017249584,\n", " -0.00035870075,\n", " 0.0001296997,\n", " 0.0010085106,\n", " -0.0015490055,\n", " 0.0011205673,\n", " 0.0005619526,\n", " 0.00039994717,\n", " 0.00010895729,\n", " -0.00030744076,\n", " -0.00088620186,\n", " 0.00041878223,\n", " -0.00022542477,\n", " -0.00011742115,\n", " 0.00050103664,\n", " -0.0016771555,\n", " 0.00026583672,\n", " 0.0017547607,\n", " -0.00030744076,\n", " -9.739399e-05,\n", " -5.733967e-05,\n", " -0.00091278553,\n", " 0.00035226345,\n", " 0.0005656481,\n", " 3.4451485e-05,\n", " 0.00035190582,\n", " 0.0002849102,\n", " -0.00033521652,\n", " -0.00017440319,\n", " 0.0005315542,\n", " -8.201599e-05,\n", " -0.0004361868,\n", " -0.00045263767,\n", " 0.00045883656,\n", " 1.4305115e-05,\n", " -0.00022399426,\n", " 0.00053572655,\n", " 1.6450882e-05,\n", " -0.00056290627,\n", " -0.0017770529,\n", " 0.0005773306,\n", " 0.00046241283,\n", " 0.0006300211,\n", " -0.00024795532,\n", " -0.00057172775,\n", " 0.00021660328,\n", " -0.0006250143,\n", " 0.0005772114,\n", " 0.0008751154,\n", " 0.00027799606,\n", " -0.0009586811,\n", " -0.0006481409,\n", " 0.0007199049,\n", " 0.00012719631,\n", " 0.00067567825,\n", " -3.862381e-05,\n", " -0.00015974045,\n", " -0.0011217594,\n", " 0.00055229664,\n", " 0.00052154064,\n", " 0.00014197826,\n", " -0.00078713894,\n", " 0.00032556057,\n", " -0.00010406971,\n", " 0.0005441904,\n", " -0.00018823147,\n", " -0.00055229664,\n", " 0.0001860857,\n", " 0.00030636787,\n", " -0.0005117655,\n", " -0.00036227703,\n", " 5.2690506e-05,\n", " 0.00095272064,\n", " 0.00016331673,\n", " -8.714199e-05,\n", " -0.0003143549,\n", " 0.00023400784,\n", " 1.04904175e-05,\n", " -0.0005246401,\n", " 0.00048089027,\n", " 0.00057446957,\n", " -0.0006082058,\n", " -0.00043809414,\n", " -0.0005866289,\n", " -0.0008265972,\n", " 0.001024723,\n", " 9.3340874e-05,\n", " -0.0018473864,\n", " 0.0012134314,\n", " 0.0007991791,\n", " 8.177757e-05,\n", " 0.00077950954,\n", " 0.000269413,\n", " 0.0002478361,\n", " 5.00679e-06,\n", " -0.00023591518,\n", " -0.00041258335,\n", " -0.0008803606,\n", " -0.0009549856,\n", " -0.0005443096,\n", " 0.0012031794,\n", " 0.00078451633,\n", " -0.000651598,\n", " 0.0007582903,\n", " -0.00027132034,\n", " -0.00015103817,\n", " 0.0010640621,\n", " -5.1617622e-05,\n", " -0.00069737434,\n", " 0.0005290508,\n", " 3.2782555e-05,\n", " -0.0006006956,\n", " 6.16312e-05,\n", " -1.5735626e-05,\n", " 0.00021910667,\n", " -0.00022625923,\n", " 0.00054860115,\n", " 0.00042176247,\n", " -0.00021910667,\n", " -0.00021779537,\n", " 3.0636787e-05,\n", " -0.0006351471,\n", " -9.536743e-06,\n", " 0.0004980564,\n", " 0.0002501011,\n", " 6.353855e-05,\n", " 9.047985e-05,\n", " -0.00020289421,\n", " -0.00043570995,\n", " -0.00049960613,\n", " 0.0002861023,\n", " 0.00064480305,\n", " 9.9778175e-05,\n", " -0.00022125244,\n", " -0.00021016598,\n", " -0.0004131794,\n", " -0.00011241436,\n", " 0.00037169456,\n", " -0.00018417835,\n", " 0.00026512146,\n", " 0.0005210638,\n", " -0.00046491623,\n", " -0.00049352646,\n", " 0.00021958351,\n", " -0.0004351139,\n", " 0.00010418892,\n", " 0.0002578497,\n", " -3.695488e-05,\n", " 8.571148e-05,\n", " 0.0006535053,\n", " 0.0002617836,\n", " -0.00044703484,\n", " -0.0004891157,\n", " 5.197525e-05,\n", " -0.00010609627,\n", " 0.00015306473,\n", " 0.00015306473,\n", " 0.00021147728,\n", " -1.41859055e-05,\n", " 4.4465065e-05,\n", " 8.749962e-05,\n", " -2.9802322e-06,\n", " -1.6570091e-05,\n", " 0.00010943413,\n", " 9.36985e-05,\n", " -0.0001654625,\n", " -0.00029182434,\n", " 6.67572e-05,\n", " -0.00022661686,\n", " -0.0017278194,\n", " -0.00023519993,\n", " 0.0017248392,\n", " 0.00046360493,\n", " -0.00024092197,\n", " -0.00045108795,\n", " 0.00019359589,\n", " 0.00024354458,\n", " 4.1604042e-05,\n", " 0.00016129017,\n", " -0.000461936,\n", " -0.00042796135,\n", " 6.9618225e-05,\n", " 0.00032639503,\n", " 0.00019872189,\n", " 0.0003451109,\n", " 0.0001360178,\n", " 4.529953e-05,\n", " 0.000106453896,\n", " -0.00029468536,\n", " -0.0005502701,\n", " -7.7962875e-05,\n", " 0.00019323826,\n", " 9.560585e-05,\n", " 0.0003848076,\n", " 0.00017344952,\n", " -0.000259161,\n", " -6.997585e-05,\n", " 0.00020575523,\n", " 7.70092e-05,\n", " -7.1525574e-05,\n", " 1.5377998e-05,\n", " 6.747246e-05,\n", " -8.571148e-05,\n", " -0.00023043156,\n", " -0.00031542778,\n", " -0.00017988682,\n", " 0.00011777878,\n", " 0.00027179718,\n", " 9.262562e-05,\n", " 3.0040741e-05,\n", " -0.00010108948,\n", " 4.1365623e-05,\n", " 0.00012946129,\n", " -0.00032234192,\n", " -0.0003246069,\n", " 0.00025570393,\n", " 0.00029075146,\n", " -8.702278e-05,\n", " -0.00041627884,\n", " -0.00042831898,\n", " 0.00031030178,\n", " 0.00047779083,\n", " 0.00016498566,\n", " -8.332729e-05,\n", " -0.0005466938,\n", " -0.00051522255,\n", " 0.00012612343,\n", " 0.00031006336,\n", " 0.00016760826,\n", " 0.00015997887,\n", " 0.00025093555,\n", " -0.00013792515,\n", " -0.00033402443,\n", " -0.0001988411,\n", " -8.046627e-05,\n", " 0.0005007982,\n", " 0.0006226301,\n", " 0.00017249584,\n", " -0.00034821033,\n", " -0.0011255741,\n", " -0.00065505505,\n", " 0.0009601116,\n", " 0.00082445145,\n", " -0.00011217594,\n", " -0.0001282692,\n", " 0.0002770424,\n", " 0.00015306473,\n", " -0.00019085407,\n", " -0.0001758337,\n", " -0.00010287762,\n", " -8.821487e-06,\n", " -0.00010919571,\n", " -0.000516057,\n", " -0.00092697144,\n", " 8.833408e-05,\n", " 0.0012649298,\n", " 0.00051259995,\n", " -0.00018763542,\n", " -6.3180923e-06,\n", " 0.00024533272,\n", " 0.00027942657,\n", " -1.4424324e-05,\n", " -0.00051021576,\n", " -0.0005968809,\n", " -0.00033593178,\n", " 6.67572e-06,\n", " 0.00019788742,\n", " 0.00023579597,\n", " -0.00016510487,\n", " -0.00011968613,\n", " 0.0006773472,\n", " 0.00054967403,\n", " 0.000120162964,\n", " 3.8981438e-05,\n", " -8.4757805e-05,\n", " -0.00036895275,\n", " -0.0006855726,\n", " -0.00033724308,\n", " 0.000400424,\n", " 0.0004400015,\n", " 0.0001603365,\n", " 6.7949295e-06,\n", " -0.00019705296,\n", " 9.059906e-06,\n", " 0.0003259182,\n", " 0.00020599365,\n", " -0.00014531612,\n", " -0.0006175041,\n", " -0.00051891804,\n", " 0.00039041042,\n", " 0.00042927265,\n", " -0.00038158894,\n", " -0.00019919872,\n", " 0.00044381618,\n", " 0.0005271435,\n", " 0.0002810955,\n", " -3.170967e-05,\n", " -0.00028276443,\n", " -0.00040602684,\n", " -0.00018298626,\n", " 0.00025832653,\n", " 0.00017535686,\n", " -0.00043952465,\n", " -0.00031387806,\n", " 0.0004760027,\n", " 0.0005055666,\n", " -3.9577484e-05,\n", " -0.0001538992,\n", " 5.054474e-05,\n", " 9.23872e-05,\n", " 2.4795532e-05,\n", " 9.202957e-05,\n", " 2.18153e-05,\n", " -0.00019145012,\n", " -0.00022768974,\n", " -7.534027e-05,\n", " 0.000116825104,\n", " 0.00013387203,\n", " -0.00014436245,\n", " -0.00035130978,\n", " -0.00020563602,\n", " -2.8133392e-05,\n", " 0.00011968613,\n", " 0.00037670135,\n", " 0.00040066242,\n", " 0.0001180172,\n", " -0.00015282631,\n", " -0.0001989603,\n", " 7.867813e-06,\n", " 0.0001386404,\n", " -6.484985e-05,\n", " -0.00045239925,\n", " -0.0005723238,\n", " -8.9764595e-05,\n", " 0.0003631115,\n", " 0.00029706955,\n", " 0.00010049343,\n", " 5.2332878e-05,\n", " 0.00015938282,\n", " 0.00017654896,\n", " -5.6028366e-06,\n", " -0.00020158291,\n", " -0.0003437996,\n", " -0.00022304058,\n", " 0.00011599064,\n", " 0.00025689602,\n", " 0.00014984608,\n", " -0.00018048286,\n", " -0.00039458275,\n", " 2.503395e-05,\n", " 0.0004348755,\n", " 0.0002207756,\n", " -0.00015580654,\n", " -0.0001386404,\n", " -1.2993813e-05,\n", " -0.00017011166,\n", " -0.0003708601,\n", " -0.0002925396,\n", " 0.0002259016,\n", " 0.0005816221,\n", " 0.00037300587,\n", " 0.00012350082,\n", " -4.6372414e-05,\n", " -0.00019288063,\n", " -0.00012278557,\n", " -1.2636185e-05,\n", " -1.0728836e-05,\n", " 3.874302e-05,\n", " 5.2332878e-05,\n", " -3.6239624e-05,\n", " -0.00021469593,\n", " -0.00040018559,\n", " -0.00022852421,\n", " 4.4226646e-05,\n", " 9.512901e-05,\n", " 0.00018906593,\n", " 0.0001629591,\n", " -4.6253204e-05,\n", " 2.9802322e-06,\n", " 0.00014805794,\n", " 6.556511e-05,\n", " -0.00014853477,\n", " -0.000115036964,\n", " 0.00022244453,\n", " 0.00034832954,\n", " 9.250641e-05,\n", " -0.00019466877,\n", " -0.0002720356,\n", " -5.2452087e-05,\n", " 0.00021719933,\n", " 0.00015330315,\n", " -7.9631805e-05,\n", " -5.8293343e-05,\n", " 0.00012660027,\n", " 0.00018894672,\n", " 9.596348e-05,\n", " -0.00021409988,\n", " -0.00072836876,\n", " -0.00070512295,\n", " 0.00012600422,\n", " 0.0007597208,\n", " 0.0005351305,\n", " 3.4332275e-05,\n", " -0.00013613701,\n", " -8.34465e-07,\n", " 3.3974648e-05,\n", " -0.00020313263,\n", " -0.00041115284,\n", " -0.00022399426,\n", " 0.00016844273,\n", " 0.00035524368,\n", " 0.00031149387,\n", " 0.00021255016,\n", " 4.553795e-05,\n", " -0.00011384487,\n", " -0.00018846989,\n", " -0.00014972687,\n", " 1.8239021e-05,\n", " 0.00014984608,\n", " 0.00011527538,\n", " -0.00010728836,\n", " -0.00031590462,\n", " -0.00014150143,\n", " 0.00025129318,\n", " 0.00034487247,\n", " 0.00016140938,\n", " -3.33786e-06,\n", " -0.00012373924,\n", " -0.00013613701,\n", " 1.1086464e-05,\n", " 8.4519386e-05,\n", " -2.4676323e-05,\n", " -6.0915947e-05,\n", " 8.022785e-05,\n", " 0.00017356873,\n", " 0.000113248825,\n", " -5.364418e-06,\n", " -0.00011265278,\n", " -0.00014686584,\n", " -7.7843666e-05,\n", " 3.325939e-05,\n", " 7.522106e-05,\n", " 3.170967e-05,\n", " -4.2915344e-06,\n", " -4.053116e-06,\n", " -0.000121593475,\n", " -0.0003927946,\n", " -0.00053858757,\n", " -0.00035703182,\n", " -3.9339066e-05,\n", " 0.00024485588,\n", " 0.00045323372,\n", " 0.0004183054,\n", " 0.00011765957,\n", " -5.2809715e-05,\n", " 3.385544e-05,\n", " 0.00012075901,\n", " 0.0001052618,\n", " 4.3153763e-05,\n", " -0.00012600422,\n", " -0.0003684759,\n", " -0.00036931038,\n", " -6.771088e-05,\n", " 9.191036e-05,\n", " -5.4240227e-05,\n", " -0.00017929077,\n", " -0.00016903877,\n", " -5.4955482e-05,\n", " 0.00018310547,\n", " 0.00035107136,\n", " 0.00022113323,\n", " -9.930134e-05,\n", " -0.00025820732,\n", " -0.00014853477,\n", " 2.002716e-05,\n", " 0.00014984608,\n", " 0.00026202202,\n", " 0.0002746582,\n", " 0.00014615059,\n", " 2.3126602e-05,\n", " -0.000104784966,\n", " -0.00032019615,\n", " -0.0003950596,\n", " -0.0001565218,\n", " 0.00015974045,\n", " 0.0001295805,\n", " -0.00020837784,\n", " -0.00026977062,\n", " 9.453297e-05,\n", " 0.0003759861,\n", " 0.00024235249,\n", " 4.887581e-06,\n", " 8.34465e-07,\n", " 0.00015199184,\n", " 0.00027418137,\n", " 0.0002901554,\n", " 0.00015771389,\n", " -9.191036e-05,\n", " -0.00024962425,\n", " -0.00017285347,\n", " 3.159046e-05,\n", " 0.000118255615,\n", " -5.2332878e-05,\n", " -0.00039553642,\n", " -0.00063848495,\n", " -0.00043666363,\n", " 0.00010180473,\n", " 0.0004720688,\n", " 0.00032770634,\n", " -6.723404e-05,\n", " -0.00017023087,\n", " 0.00011622906,\n", " 0.00042903423,\n", " 0.00040853024,\n", " 1.9788742e-05,\n", " -0.00043725967,\n", " -0.00054872036,\n", " -0.00020849705,\n", " 0.0002565384,\n", " 0.00044894218,\n", " 0.00026988983,\n", " -6.0796738e-06,\n", " -0.00023794174,\n", " -0.0003335476,\n", " -0.00016093254,\n", " 0.00014710426,\n", " 0.0003414154,\n", " 0.00031983852,\n", " 0.00022983551,\n", " 0.00019466877,\n", " 9.7990036e-05,\n", " -0.00013780594,\n", " -0.0003979206,\n", " -0.00048244,\n", " -0.0002809763,\n", " 5.340576e-05,\n", " 0.00022733212,\n", " 0.00010108948,\n", " -0.00011706352,\n", " -0.00017797947,\n", " -0.00010204315,\n", " -3.874302e-05,\n", " -9.3340874e-05,\n", " -0.0002501011,\n", " -0.00038826466,\n", " -0.00034701824,\n", " -4.9829483e-05,\n", " 0.00033414364,\n", " 0.00053167343,\n", " 0.0005091429,\n", " 0.00040447712,\n", " 0.00035309792,\n", " 0.00033426285,\n", " 0.0002733469,\n", " 0.00016772747,\n", " 5.1021576e-05,\n", " -4.2676926e-05,\n", " -0.00016117096,\n", " -0.00024998188,\n", " -0.00023853779,\n", " -0.00018048286,\n", " -0.00014603138,\n", " -0.00016987324,\n", " -0.00018107891,\n", " -0.00012433529,\n", " -2.5629997e-05,\n", " 3.1471252e-05,\n", " 3.4451485e-05,\n", " 5.4001808e-05,\n", " 9.441376e-05,\n", " 0.00013113022,\n", " 0.00012397766,\n", " 8.440018e-05,\n", " 6.318092e-05,\n", " 4.5776367e-05,\n", " 1.8119812e-05,\n", " -1.1920929e-06,\n", " 9.655952e-06,\n", " 2.7179718e-05,\n", " 8.106232e-06,\n", " -3.9219856e-05,\n", " -7.1287155e-05,\n", " -5.9962273e-05,\n", " -1.5377998e-05,\n", " 3.0517578e-05,\n", " 4.4703484e-05,\n", " 4.4107437e-06,\n", " -6.389618e-05,\n", " -9.930134e-05,\n", " -7.3194504e-05,\n", " -1.2874603e-05,\n", " 3.8146973e-05,\n", " 5.6624413e-05,\n", " 5.531311e-05,\n", " 4.5895576e-05,\n", " 1.1444092e-05,\n", " -5.4001808e-05,\n", " -0.00016379356,\n", " -0.00033259392,\n", " -0.0004862547,\n", " -0.00048577785,\n", " -0.0002746582,\n", " 5.877018e-05,\n", " 0.00035071373,\n", " 0.00048100948,\n", " 0.000461936,\n", " 0.0003658533,\n", " 0.0002297163,\n", " 7.1406364e-05,\n", " -6.723404e-05,\n", " -0.00014436245,\n", " -0.00015163422,\n", " -0.00011277199,\n", " -6.7830086e-05,\n", " -4.0769577e-05,\n", " -1.335144e-05,\n", " 2.7537346e-05,\n", " 4.9829483e-05,\n", " 8.34465e-06,\n", " -9.679794e-05,\n", " -0.000187397,\n", " -0.00017368793,\n", " -5.698204e-05,\n", " 8.380413e-05,\n", " 0.00017786026,\n", " 0.00020968914,\n", " 0.00019919872,\n", " 0.00016152859,\n", " 9.512901e-05,\n", " -2.503395e-06,\n", " -0.00011241436,\n", " -0.00020515919,\n", " -0.0002411604,\n", " -0.0001860857,\n", " -4.720688e-05,\n", " 0.00011241436,\n", " 0.00020778179,\n", " 0.00018501282,\n", " 6.055832e-05,\n", " -8.761883e-05,\n", " -0.00019109249,\n", " -0.00022399426,\n", " -0.00017642975,\n", " -6.723404e-05,\n", " 4.7683716e-05,\n", " 0.00010752678,\n", " 7.70092e-05,\n", " -4.196167e-05,\n", " -0.00019216537,\n", " -0.0002835989,\n", " -0.0002515316,\n", " -8.8214874e-05,\n", " 0.00013697147,\n", " 0.0003067255,\n", " 0.00034451485,\n", " 0.00026738644,\n", " ...],\n", " 'likelihood.second_noise_covar.noise': [array([0.36738268], dtype=float32),\n", " array([0.38573632], dtype=float32),\n", " array([0.40315896], dtype=float32),\n", " array([0.41900876], dtype=float32),\n", " array([0.43215463], dtype=float32),\n", " array([0.44077873], dtype=float32),\n", " array([0.44490042], dtype=float32),\n", " array([0.44583604], dtype=float32),\n", " array([0.4441356], dtype=float32),\n", " array([0.4402297], dtype=float32),\n", " array([0.43475932], dtype=float32),\n", " array([0.42851534], dtype=float32),\n", " array([0.4219687], dtype=float32),\n", " array([0.41550168], dtype=float32),\n", " array([0.40991175], dtype=float32),\n", " array([0.40548086], dtype=float32),\n", " array([0.40268824], dtype=float32),\n", " array([0.4017722], dtype=float32),\n", " array([0.40262595], dtype=float32),\n", " array([0.40552768], dtype=float32),\n", " array([0.41011485], dtype=float32),\n", " array([0.41543463], dtype=float32),\n", " array([0.4204038], dtype=float32),\n", " array([0.42421585], dtype=float32),\n", " array([0.42780086], dtype=float32),\n", " array([0.43122038], dtype=float32),\n", " array([0.4342048], dtype=float32),\n", " array([0.43633205], dtype=float32),\n", " array([0.43764648], dtype=float32),\n", " array([0.43837819], dtype=float32),\n", " array([0.43871656], dtype=float32),\n", " array([0.43889502], dtype=float32),\n", " array([0.43830147], dtype=float32),\n", " array([0.43679112], dtype=float32),\n", " array([0.43502226], dtype=float32),\n", " array([0.43307117], dtype=float32),\n", " array([0.43100366], dtype=float32),\n", " array([0.4285694], dtype=float32),\n", " array([0.42614648], dtype=float32),\n", " array([0.42403382], dtype=float32),\n", " array([0.4228659], dtype=float32),\n", " array([0.42196563], dtype=float32),\n", " array([0.421649], dtype=float32),\n", " array([0.42192906], dtype=float32),\n", " array([0.4232269], dtype=float32),\n", " array([0.42542967], dtype=float32),\n", " array([0.42811984], dtype=float32),\n", " array([0.43092394], dtype=float32),\n", " array([0.4333397], dtype=float32),\n", " array([0.4353848], dtype=float32),\n", " array([0.43702033], dtype=float32),\n", " array([0.43836138], dtype=float32),\n", " array([0.43853486], dtype=float32),\n", " array([0.43773994], dtype=float32),\n", " array([0.435543], dtype=float32),\n", " array([0.4329918], dtype=float32),\n", " array([0.43078515], dtype=float32),\n", " array([0.42896307], dtype=float32),\n", " array([0.42769653], dtype=float32),\n", " array([0.42688942], dtype=float32),\n", " array([0.42652223], dtype=float32),\n", " array([0.4270559], dtype=float32),\n", " array([0.42780793], dtype=float32),\n", " array([0.42895147], dtype=float32),\n", " array([0.43034196], dtype=float32),\n", " array([0.43138093], dtype=float32),\n", " array([0.43203393], dtype=float32),\n", " array([0.4325035], dtype=float32),\n", " array([0.4326807], dtype=float32),\n", " array([0.43225908], dtype=float32),\n", " array([0.43137145], dtype=float32),\n", " array([0.4304383], dtype=float32),\n", " array([0.42990115], dtype=float32),\n", " array([0.42971265], dtype=float32),\n", " array([0.42958096], dtype=float32),\n", " array([0.42949793], dtype=float32),\n", " array([0.42958838], dtype=float32),\n", " array([0.43001065], dtype=float32),\n", " array([0.42977962], dtype=float32),\n", " array([0.42909306], dtype=float32),\n", " array([0.42882892], dtype=float32),\n", " array([0.42848876], dtype=float32),\n", " array([0.4287791], dtype=float32),\n", " array([0.42952102], dtype=float32),\n", " array([0.4306578], dtype=float32),\n", " array([0.43162912], dtype=float32),\n", " array([0.43222916], dtype=float32),\n", " array([0.43284526], dtype=float32),\n", " array([0.43312725], dtype=float32),\n", " array([0.43332925], dtype=float32),\n", " array([0.4324531], dtype=float32),\n", " array([0.43165454], dtype=float32),\n", " array([0.4307673], dtype=float32),\n", " array([0.43031168], dtype=float32),\n", " array([0.42972505], dtype=float32),\n", " array([0.42939553], dtype=float32),\n", " array([0.42867577], dtype=float32),\n", " array([0.42766252], dtype=float32),\n", " array([0.4270329], dtype=float32),\n", " array([0.42690432], dtype=float32),\n", " array([0.42743835], dtype=float32),\n", " array([0.42828196], dtype=float32),\n", " array([0.42946166], dtype=float32),\n", " array([0.4309412], dtype=float32),\n", " array([0.43254954], dtype=float32),\n", " array([0.43320528], dtype=float32),\n", " array([0.43310344], dtype=float32),\n", " array([0.43284556], dtype=float32),\n", " array([0.43252316], dtype=float32),\n", " array([0.43183357], dtype=float32),\n", " array([0.43139663], dtype=float32),\n", " array([0.43122813], dtype=float32),\n", " array([0.43103334], dtype=float32),\n", " array([0.43094698], dtype=float32),\n", " array([0.43069652], dtype=float32),\n", " array([0.43056884], dtype=float32),\n", " array([0.43070704], dtype=float32),\n", " array([0.43110612], dtype=float32),\n", " array([0.43154982], dtype=float32),\n", " array([0.43214104], dtype=float32),\n", " array([0.43266743], dtype=float32),\n", " array([0.4329447], dtype=float32),\n", " array([0.43285245], dtype=float32),\n", " array([0.43236506], dtype=float32),\n", " array([0.4319885], dtype=float32),\n", " array([0.43164915], dtype=float32),\n", " array([0.4314433], dtype=float32),\n", " array([0.43140963], dtype=float32),\n", " array([0.43150935], dtype=float32),\n", " array([0.43165725], dtype=float32),\n", " array([0.43165335], dtype=float32),\n", " array([0.43137783], dtype=float32),\n", " array([0.4306439], dtype=float32),\n", " array([0.4302013], dtype=float32),\n", " array([0.4295596], dtype=float32),\n", " array([0.42936063], dtype=float32),\n", " array([0.42927024], dtype=float32),\n", " array([0.42972916], dtype=float32),\n", " array([0.4305859], dtype=float32),\n", " array([0.43067014], dtype=float32),\n", " array([0.43078429], dtype=float32),\n", " array([0.43095377], dtype=float32),\n", " array([0.43100825], dtype=float32),\n", " array([0.43122146], dtype=float32),\n", " array([0.4314861], dtype=float32),\n", " array([0.43144658], dtype=float32),\n", " array([0.43123612], dtype=float32),\n", " array([0.43089938], dtype=float32),\n", " array([0.43017277], dtype=float32),\n", " array([0.42979783], dtype=float32),\n", " array([0.42993212], dtype=float32),\n", " array([0.42966583], dtype=float32),\n", " array([0.42955697], dtype=float32),\n", " array([0.42974362], dtype=float32),\n", " array([0.4301831], dtype=float32),\n", " array([0.43069336], dtype=float32),\n", " array([0.43095353], dtype=float32),\n", " array([0.4310313], dtype=float32),\n", " array([0.43127635], dtype=float32),\n", " array([0.43155152], dtype=float32),\n", " array([0.43182144], dtype=float32),\n", " array([0.4318326], dtype=float32),\n", " array([0.4318232], dtype=float32),\n", " array([0.4312838], dtype=float32),\n", " array([0.43088016], dtype=float32),\n", " array([0.4307506], dtype=float32),\n", " array([0.43090072], dtype=float32),\n", " array([0.4304179], dtype=float32),\n", " array([0.4300452], dtype=float32),\n", " array([0.42970324], dtype=float32),\n", " array([0.42940482], dtype=float32),\n", " array([0.42943627], dtype=float32),\n", " array([0.4298144], dtype=float32),\n", " array([0.43036577], dtype=float32),\n", " array([0.4309442], dtype=float32),\n", " array([0.43139234], dtype=float32),\n", " array([0.4313382], dtype=float32),\n", " array([0.43071005], dtype=float32),\n", " array([0.4303153], dtype=float32),\n", " array([0.4292168], dtype=float32),\n", " array([0.42846796], dtype=float32),\n", " array([0.4283507], dtype=float32),\n", " array([0.42870143], dtype=float32),\n", " array([0.4293788], dtype=float32),\n", " array([0.43023944], dtype=float32),\n", " array([0.43096536], dtype=float32),\n", " array([0.4317678], dtype=float32),\n", " array([0.43220928], dtype=float32),\n", " array([0.43259713], dtype=float32),\n", " array([0.43243057], dtype=float32),\n", " array([0.43181908], dtype=float32),\n", " array([0.4307251], dtype=float32),\n", " array([0.42998102], dtype=float32),\n", " array([0.4296319], dtype=float32),\n", " array([0.42964536], dtype=float32),\n", " array([0.42983317], dtype=float32),\n", " array([0.43032363], dtype=float32),\n", " array([0.43077916], dtype=float32),\n", " array([0.43129838], dtype=float32),\n", " array([0.43170217], dtype=float32),\n", " array([0.43205407], dtype=float32),\n", " array([0.43157735], dtype=float32),\n", " array([0.43076637], dtype=float32),\n", " array([0.42980593], dtype=float32),\n", " array([0.4292232], dtype=float32),\n", " array([0.4290947], dtype=float32),\n", " array([0.42897764], dtype=float32),\n", " array([0.42945325], dtype=float32),\n", " array([0.43040514], dtype=float32),\n", " array([0.43162927], dtype=float32),\n", " array([0.43283665], dtype=float32),\n", " array([0.43368298], dtype=float32),\n", " array([0.43377078], dtype=float32),\n", " array([0.43302554], dtype=float32),\n", " array([0.4321506], dtype=float32),\n", " array([0.43100852], dtype=float32),\n", " array([0.42961383], dtype=float32),\n", " array([0.42788374], dtype=float32),\n", " array([0.42691123], dtype=float32),\n", " array([0.4264415], dtype=float32),\n", " array([0.42691052], dtype=float32),\n", " array([0.4280402], dtype=float32),\n", " array([0.4290501], dtype=float32),\n", " array([0.43024662], dtype=float32),\n", " array([0.43132383], dtype=float32),\n", " array([0.43235198], dtype=float32),\n", " array([0.43305132], dtype=float32),\n", " array([0.4331415], dtype=float32),\n", " array([0.43258107], dtype=float32),\n", " array([0.43160936], dtype=float32),\n", " array([0.43056634], dtype=float32),\n", " array([0.42904449], dtype=float32),\n", " array([0.42745066], dtype=float32),\n", " array([0.42686203], dtype=float32),\n", " array([0.42743745], dtype=float32),\n", " array([0.42868072], dtype=float32),\n", " array([0.4304234], dtype=float32),\n", " array([0.43219587], dtype=float32),\n", " array([0.4335505], dtype=float32),\n", " array([0.43334576], dtype=float32),\n", " array([0.4314322], dtype=float32),\n", " array([0.42916492], dtype=float32),\n", " array([0.42718855], dtype=float32),\n", " array([0.42626944], dtype=float32),\n", " array([0.42657116], dtype=float32),\n", " array([0.42805547], dtype=float32),\n", " array([0.42926115], dtype=float32),\n", " array([0.43051192], dtype=float32),\n", " array([0.431791], dtype=float32),\n", " array([0.43295747], dtype=float32),\n", " array([0.43378803], dtype=float32),\n", " array([0.4338881], dtype=float32),\n", " array([0.43298304], dtype=float32),\n", " array([0.4315181], dtype=float32),\n", " array([0.42991087], dtype=float32),\n", " array([0.4287777], dtype=float32),\n", " array([0.42759264], dtype=float32),\n", " array([0.42627037], dtype=float32),\n", " array([0.42608336], dtype=float32),\n", " array([0.42693457], dtype=float32),\n", " array([0.4286323], dtype=float32),\n", " array([0.43076584], dtype=float32),\n", " array([0.43236756], dtype=float32),\n", " array([0.43357942], dtype=float32),\n", " array([0.4342221], dtype=float32),\n", " array([0.43345344], dtype=float32),\n", " array([0.43150496], dtype=float32),\n", " array([0.42936182], dtype=float32),\n", " array([0.42706525], dtype=float32),\n", " array([0.42567602], dtype=float32),\n", " array([0.42567262], dtype=float32),\n", " array([0.42699707], dtype=float32),\n", " array([0.42932078], dtype=float32),\n", " array([0.4317068], dtype=float32),\n", " array([0.43383127], dtype=float32),\n", " array([0.43489304], dtype=float32),\n", " array([0.43487543], dtype=float32),\n", " array([0.43425697], dtype=float32),\n", " array([0.43296054], dtype=float32),\n", " array([0.43159172], dtype=float32),\n", " array([0.4300765], dtype=float32),\n", " array([0.42823106], dtype=float32),\n", " array([0.42727268], dtype=float32),\n", " array([0.42639226], dtype=float32),\n", " array([0.4261284], dtype=float32),\n", " array([0.42684045], dtype=float32),\n", " array([0.42868668], dtype=float32),\n", " array([0.43121377], dtype=float32),\n", " array([0.43347588], dtype=float32),\n", " array([0.4347134], dtype=float32),\n", " array([0.4345507], dtype=float32),\n", " array([0.43323764], dtype=float32),\n", " array([0.4313671], dtype=float32),\n", " array([0.42887026], dtype=float32),\n", " array([0.42599264], dtype=float32),\n", " array([0.42456457], dtype=float32),\n", " array([0.42492998], dtype=float32),\n", " array([0.42653668], dtype=float32),\n", " array([0.4290223], dtype=float32),\n", " array([0.43108818], dtype=float32),\n", " array([0.43245202], dtype=float32),\n", " array([0.4334743], dtype=float32),\n", " array([0.43356606], dtype=float32),\n", " array([0.43286744], dtype=float32),\n", " array([0.43128783], dtype=float32),\n", " array([0.42943767], dtype=float32),\n", " array([0.42818442], dtype=float32),\n", " array([0.42789978], dtype=float32),\n", " array([0.42806315], dtype=float32),\n", " array([0.42830348], dtype=float32),\n", " array([0.42843902], dtype=float32),\n", " array([0.42861432], dtype=float32),\n", " array([0.42939454], dtype=float32),\n", " array([0.4302188], dtype=float32),\n", " array([0.43121648], dtype=float32),\n", " array([0.4321191], dtype=float32),\n", " array([0.43284684], dtype=float32),\n", " array([0.4330524], dtype=float32),\n", " array([0.43262458], dtype=float32),\n", " array([0.43184912], dtype=float32),\n", " array([0.43107057], dtype=float32),\n", " array([0.430376], dtype=float32),\n", " array([0.42947394], dtype=float32),\n", " array([0.42910913], dtype=float32),\n", " array([0.42917198], dtype=float32),\n", " array([0.42956123], dtype=float32),\n", " array([0.43007278], dtype=float32),\n", " array([0.4304065], dtype=float32),\n", " array([0.43085474], dtype=float32),\n", " array([0.43131292], dtype=float32),\n", " array([0.4311974], dtype=float32),\n", " array([0.4308786], dtype=float32),\n", " array([0.43065727], dtype=float32),\n", " array([0.4307272], dtype=float32),\n", " array([0.43096974], dtype=float32),\n", " array([0.43115637], dtype=float32),\n", " array([0.43106496], dtype=float32),\n", " array([0.42960557], dtype=float32),\n", " array([0.42844978], dtype=float32),\n", " array([0.42827338], dtype=float32),\n", " array([0.42900738], dtype=float32),\n", " array([0.43002766], dtype=float32),\n", " array([0.43129927], dtype=float32),\n", " array([0.4321393], dtype=float32),\n", " array([0.43263558], dtype=float32),\n", " array([0.43221432], dtype=float32),\n", " array([0.43114614], dtype=float32),\n", " array([0.42958805], dtype=float32),\n", " array([0.4280429], dtype=float32),\n", " array([0.42735898], dtype=float32),\n", " array([0.42694458], dtype=float32),\n", " array([0.42737278], dtype=float32),\n", " array([0.42871097], dtype=float32),\n", " array([0.43071556], dtype=float32),\n", " array([0.43286696], dtype=float32),\n", " array([0.43446204], dtype=float32),\n", " array([0.43476576], dtype=float32),\n", " array([0.43399543], dtype=float32),\n", " array([0.43236738], dtype=float32),\n", " array([0.4303335], dtype=float32),\n", " array([0.42863464], dtype=float32),\n", " array([0.42685053], dtype=float32),\n", " array([0.4255997], dtype=float32),\n", " array([0.4260378], dtype=float32),\n", " array([0.4277156], dtype=float32),\n", " array([0.430027], dtype=float32),\n", " array([0.43226966], dtype=float32),\n", " array([0.4334591], dtype=float32),\n", " array([0.43357298], dtype=float32),\n", " array([0.43295288], dtype=float32),\n", " array([0.43185532], dtype=float32),\n", " array([0.43076527], dtype=float32),\n", " array([0.43010816], dtype=float32),\n", " array([0.42984623], dtype=float32),\n", " array([0.4299382], dtype=float32),\n", " array([0.43057975], dtype=float32),\n", " array([0.43151015], dtype=float32),\n", " array([0.4322374], dtype=float32),\n", " array([0.43239707], dtype=float32),\n", " array([0.43221915], dtype=float32),\n", " array([0.4317808], dtype=float32),\n", " array([0.43111703], dtype=float32),\n", " array([0.43069276], dtype=float32),\n", " array([0.43061715], dtype=float32),\n", " array([0.4305794], dtype=float32),\n", " array([0.42972475], dtype=float32),\n", " array([0.42862982], dtype=float32),\n", " array([0.42785698], dtype=float32),\n", " array([0.42790043], dtype=float32),\n", " array([0.42852482], dtype=float32),\n", " array([0.42921966], dtype=float32),\n", " array([0.42990386], dtype=float32),\n", " array([0.4300653], dtype=float32),\n", " array([0.43001273], dtype=float32),\n", " array([0.43023536], dtype=float32),\n", " array([0.43078682], dtype=float32),\n", " array([0.43099424], dtype=float32),\n", " array([0.43051913], dtype=float32),\n", " array([0.4298989], dtype=float32),\n", " array([0.4293984], dtype=float32),\n", " array([0.4294695], dtype=float32),\n", " array([0.43002343], dtype=float32),\n", " array([0.43076742], dtype=float32),\n", " array([0.43095618], dtype=float32),\n", " array([0.43084598], dtype=float32),\n", " array([0.43077302], dtype=float32),\n", " array([0.43079603], dtype=float32),\n", " array([0.43049532], dtype=float32),\n", " array([0.4301387], dtype=float32),\n", " array([0.42980236], dtype=float32),\n", " array([0.42984465], dtype=float32),\n", " array([0.4300988], dtype=float32),\n", " array([0.4301805], dtype=float32),\n", " array([0.4302032], dtype=float32),\n", " array([0.4303234], dtype=float32),\n", " array([0.43025905], dtype=float32),\n", " array([0.42983595], dtype=float32),\n", " array([0.42924973], dtype=float32),\n", " array([0.42917362], dtype=float32),\n", " array([0.42967087], dtype=float32),\n", " array([0.43047982], dtype=float32),\n", " array([0.43118417], dtype=float32),\n", " array([0.43172845], dtype=float32),\n", " array([0.43197992], dtype=float32),\n", " array([0.43162125], dtype=float32),\n", " array([0.43105558], dtype=float32),\n", " array([0.4307748], dtype=float32),\n", " array([0.43050617], dtype=float32),\n", " array([0.43008545], dtype=float32),\n", " array([0.4293266], dtype=float32),\n", " array([0.42803758], dtype=float32),\n", " array([0.42716843], dtype=float32),\n", " array([0.42696536], dtype=float32),\n", " array([0.42646503], dtype=float32),\n", " array([0.4264418], dtype=float32),\n", " array([0.42730936], dtype=float32),\n", " array([0.4287715], dtype=float32),\n", " array([0.43076944], dtype=float32),\n", " array([0.4328421], dtype=float32),\n", " array([0.43451834], dtype=float32),\n", " array([0.43535915], dtype=float32),\n", " array([0.43509376], dtype=float32),\n", " array([0.4336757], dtype=float32),\n", " array([0.43113342], dtype=float32),\n", " array([0.42775372], dtype=float32),\n", " array([0.4242929], dtype=float32),\n", " array([0.42245588], dtype=float32),\n", " array([0.42306814], dtype=float32),\n", " array([0.42537594], dtype=float32),\n", " array([0.4288893], dtype=float32),\n", " array([0.4323103], dtype=float32),\n", " array([0.4345824], dtype=float32),\n", " array([0.43574613], dtype=float32),\n", " array([0.43555692], dtype=float32),\n", " array([0.43383285], dtype=float32),\n", " array([0.4314777], dtype=float32),\n", " array([0.42926666], dtype=float32),\n", " array([0.42751914], dtype=float32),\n", " array([0.42669675], dtype=float32),\n", " array([0.42696244], dtype=float32),\n", " array([0.42824084], dtype=float32),\n", " array([0.4299682], dtype=float32),\n", " array([0.43187368], dtype=float32),\n", " array([0.43362013], dtype=float32),\n", " array([0.43460485], dtype=float32),\n", " array([0.4345092], dtype=float32),\n", " array([0.4334595], dtype=float32),\n", " array([0.43152368], dtype=float32),\n", " array([0.42934006], dtype=float32),\n", " array([0.4278558], dtype=float32),\n", " array([0.42759076], dtype=float32),\n", " array([0.42854422], dtype=float32),\n", " array([0.43035465], dtype=float32),\n", " array([0.4322852], dtype=float32),\n", " array([0.4335254], dtype=float32),\n", " array([0.43353373], dtype=float32),\n", " array([0.43257532], dtype=float32),\n", " array([0.4313823], dtype=float32),\n", " array([0.43038544], dtype=float32),\n", " array([0.42974576], dtype=float32),\n", " array([0.42950585], dtype=float32),\n", " array([0.42944986], dtype=float32),\n", " array([0.42951497], dtype=float32),\n", " array([0.4298758], dtype=float32),\n", " array([0.4303032], dtype=float32),\n", " array([0.43081647], dtype=float32),\n", " array([0.43155402], dtype=float32),\n", " array([0.4320244], dtype=float32),\n", " array([0.4318592], dtype=float32),\n", " array([0.43126327], dtype=float32),\n", " array([0.43020225], dtype=float32),\n", " array([0.4290866], dtype=float32),\n", " array([0.4284312], dtype=float32),\n", " array([0.42836323], dtype=float32),\n", " array([0.42890728], dtype=float32),\n", " array([0.43021315], dtype=float32),\n", " array([0.4319619], dtype=float32),\n", " array([0.43330634], dtype=float32),\n", " array([0.4336037], dtype=float32),\n", " array([0.4328891], dtype=float32),\n", " array([0.43143395], dtype=float32),\n", " array([0.42984715], dtype=float32),\n", " array([0.42873743], dtype=float32),\n", " array([0.42851904], dtype=float32),\n", " array([0.4291623], dtype=float32),\n", " array([0.43040505], dtype=float32),\n", " array([0.43185127], dtype=float32),\n", " array([0.43303472], dtype=float32),\n", " array([0.4336174], dtype=float32),\n", " array([0.43355638], dtype=float32),\n", " array([0.43299615], dtype=float32),\n", " array([0.43206015], dtype=float32),\n", " array([0.4309092], dtype=float32),\n", " array([0.4299546], dtype=float32),\n", " array([0.42931986], dtype=float32),\n", " array([0.4281277], dtype=float32),\n", " array([0.42666978], dtype=float32),\n", " array([0.4264767], dtype=float32),\n", " array([0.4277814], dtype=float32),\n", " array([0.42985925], dtype=float32),\n", " array([0.43171102], dtype=float32),\n", " array([0.432897], dtype=float32),\n", " array([0.43327284], dtype=float32),\n", " array([0.4328321], dtype=float32),\n", " array([0.4318725], dtype=float32),\n", " array([0.43047294], dtype=float32),\n", " array([0.4288471], dtype=float32),\n", " array([0.42757407], dtype=float32),\n", " array([0.42723116], dtype=float32),\n", " array([0.42798105], dtype=float32),\n", " array([0.42966944], dtype=float32),\n", " array([0.4317431], dtype=float32),\n", " array([0.43353137], dtype=float32),\n", " array([0.4345624], dtype=float32),\n", " array([0.4344187], dtype=float32),\n", " array([0.43292585], dtype=float32),\n", " array([0.43061852], dtype=float32),\n", " array([0.42842335], dtype=float32),\n", " array([0.42711192], dtype=float32),\n", " array([0.42725846], dtype=float32),\n", " array([0.42877397], dtype=float32),\n", " array([0.43087092], dtype=float32),\n", " array([0.43277827], dtype=float32),\n", " array([0.43404225], dtype=float32),\n", " array([0.43437612], dtype=float32),\n", " array([0.43372783], dtype=float32),\n", " array([0.4324061], dtype=float32),\n", " array([0.43093362], dtype=float32),\n", " array([0.42974177], dtype=float32),\n", " array([0.42904058], dtype=float32),\n", " array([0.42882258], dtype=float32),\n", " array([0.42901066], dtype=float32),\n", " array([0.42958465], dtype=float32),\n", " array([0.4304803], dtype=float32),\n", " array([0.43143263], dtype=float32),\n", " array([0.43214235], dtype=float32),\n", " array([0.43234998], dtype=float32),\n", " array([0.4320649], dtype=float32),\n", " array([0.4315029], dtype=float32),\n", " array([0.43066555], dtype=float32),\n", " array([0.42964923], dtype=float32),\n", " array([0.4289615], dtype=float32),\n", " array([0.42898604], dtype=float32),\n", " array([0.4295969], dtype=float32),\n", " array([0.43026644], dtype=float32),\n", " array([0.43050584], dtype=float32),\n", " array([0.43046036], dtype=float32),\n", " array([0.4304846], dtype=float32),\n", " array([0.43067282], dtype=float32),\n", " array([0.43088958], dtype=float32),\n", " array([0.43070513], dtype=float32),\n", " array([0.42989314], dtype=float32),\n", " array([0.42885262], dtype=float32),\n", " array([0.42814836], dtype=float32),\n", " array([0.4280867], dtype=float32),\n", " array([0.42872593], dtype=float32),\n", " array([0.42993715], dtype=float32),\n", " array([0.43117085], dtype=float32),\n", " array([0.4318137], dtype=float32),\n", " array([0.43158966], dtype=float32),\n", " array([0.430612], dtype=float32),\n", " array([0.4295964], dtype=float32),\n", " array([0.42928067], dtype=float32),\n", " array([0.42980415], dtype=float32),\n", " array([0.43067795], dtype=float32),\n", " array([0.4308436], dtype=float32),\n", " array([0.42989734], dtype=float32),\n", " array([0.42885545], dtype=float32),\n", " array([0.42862082], dtype=float32),\n", " array([0.42912665], dtype=float32),\n", " array([0.43004388], dtype=float32),\n", " array([0.43119675], dtype=float32),\n", " array([0.43227607], dtype=float32),\n", " array([0.43280852], dtype=float32),\n", " array([0.43255192], dtype=float32),\n", " array([0.4316083], dtype=float32),\n", " array([0.43035188], dtype=float32),\n", " array([0.42915717], dtype=float32),\n", " array([0.4280544], dtype=float32),\n", " array([0.42678654], dtype=float32),\n", " array([0.42586944], dtype=float32),\n", " array([0.42642415], dtype=float32),\n", " array([0.42847404], dtype=float32),\n", " array([0.4310336], dtype=float32),\n", " array([0.43315464], dtype=float32),\n", " array([0.43432647], dtype=float32),\n", " array([0.43444374], dtype=float32),\n", " array([0.4335753], dtype=float32),\n", " array([0.43179035], dtype=float32),\n", " array([0.42938748], dtype=float32),\n", " array([0.42701256], dtype=float32),\n", " array([0.42547733], dtype=float32),\n", " array([0.42538264], dtype=float32),\n", " array([0.42676416], dtype=float32),\n", " array([0.428857], dtype=float32),\n", " array([0.4307852], dtype=float32),\n", " array([0.43233028], dtype=float32),\n", " array([0.43338656], dtype=float32),\n", " array([0.4337297], dtype=float32),\n", " array([0.43333355], dtype=float32),\n", " array([0.43234256], dtype=float32),\n", " array([0.4309045], dtype=float32),\n", " array([0.4290983], dtype=float32),\n", " array([0.42736983], dtype=float32),\n", " array([0.42656997], dtype=float32),\n", " array([0.42720878], dtype=float32),\n", " array([0.4290276], dtype=float32),\n", " array([0.4312617], dtype=float32),\n", " array([0.43296456], dtype=float32),\n", " array([0.4336047], dtype=float32),\n", " array([0.43327346], dtype=float32),\n", " array([0.43232197], dtype=float32),\n", " array([0.43104774], dtype=float32),\n", " array([0.4295131], dtype=float32),\n", " array([0.42795047], dtype=float32),\n", " array([0.42714986], dtype=float32),\n", " array([0.42760608], dtype=float32),\n", " array([0.42880413], dtype=float32),\n", " array([0.43010944], dtype=float32),\n", " array([0.431332], dtype=float32),\n", " array([0.43240908], dtype=float32),\n", " array([0.43316764], dtype=float32),\n", " array([0.43336305], dtype=float32),\n", " array([0.4328042], dtype=float32),\n", " array([0.4314892], dtype=float32),\n", " array([0.4298268], dtype=float32),\n", " array([0.4285992], dtype=float32),\n", " array([0.42831248], dtype=float32),\n", " array([0.4286851], dtype=float32),\n", " array([0.42932883], dtype=float32),\n", " array([0.43029878], dtype=float32),\n", " array([0.43156996], dtype=float32),\n", " array([0.43264908], dtype=float32),\n", " array([0.4330779], dtype=float32),\n", " array([0.43279693], dtype=float32),\n", " array([0.43203077], dtype=float32),\n", " array([0.43111542], dtype=float32),\n", " array([0.43045565], dtype=float32),\n", " array([0.43028694], dtype=float32),\n", " array([0.4305038], dtype=float32),\n", " array([0.43083376], dtype=float32),\n", " array([0.43109435], dtype=float32),\n", " array([0.43127874], dtype=float32),\n", " array([0.43142486], dtype=float32),\n", " array([0.43139282], dtype=float32),\n", " array([0.43097383], dtype=float32),\n", " array([0.43021438], dtype=float32),\n", " array([0.42940435], dtype=float32),\n", " array([0.42893314], dtype=float32),\n", " array([0.4291998], dtype=float32),\n", " array([0.43032014], dtype=float32),\n", " array([0.4318965], dtype=float32),\n", " array([0.43321818], dtype=float32),\n", " array([0.43370846], dtype=float32),\n", " array([0.43327942], dtype=float32),\n", " array([0.43226472], dtype=float32),\n", " array([0.43103173], dtype=float32),\n", " array([0.4297187], dtype=float32),\n", " array([0.42841133], dtype=float32),\n", " array([0.4275353], dtype=float32),\n", " array([0.4276346], dtype=float32),\n", " array([0.42880064], dtype=float32),\n", " array([0.43056604], dtype=float32),\n", " array([0.4322644], dtype=float32),\n", " array([0.43339574], dtype=float32),\n", " array([0.43373305], dtype=float32),\n", " array([0.43323633], dtype=float32),\n", " array([0.432029], dtype=float32),\n", " array([0.4303698], dtype=float32),\n", " array([0.42874274], dtype=float32),\n", " array([0.4278122], dtype=float32),\n", " array([0.4280404], dtype=float32),\n", " array([0.42933944], dtype=float32),\n", " array([0.43099752], dtype=float32),\n", " array([0.43210497], dtype=float32),\n", " array([0.43232635], dtype=float32),\n", " array([0.4319664], dtype=float32),\n", " array([0.4313712], dtype=float32),\n", " array([0.43067488], dtype=float32),\n", " array([0.43004543], dtype=float32),\n", " array([0.42969444], dtype=float32),\n", " array([0.42960277], dtype=float32),\n", " array([0.42950705], dtype=float32),\n", " array([0.42923525], dtype=float32),\n", " array([0.4290698], dtype=float32),\n", " array([0.429477], dtype=float32),\n", " array([0.4305251], dtype=float32),\n", " array([0.43184784], dtype=float32),\n", " array([0.43289143], dtype=float32),\n", " array([0.43316576], dtype=float32),\n", " array([0.43255752], dtype=float32),\n", " array([0.43135974], dtype=float32),\n", " array([0.43006632], dtype=float32),\n", " array([0.42920175], dtype=float32),\n", " array([0.42909318], dtype=float32),\n", " array([0.4296784], dtype=float32),\n", " array([0.43049747], dtype=float32),\n", " array([0.43092617], dtype=float32),\n", " array([0.43068188], dtype=float32),\n", " array([0.42995292], dtype=float32),\n", " array([0.4291392], dtype=float32),\n", " array([0.42869127], dtype=float32),\n", " array([0.4288499], dtype=float32),\n", " array([0.4294576], dtype=float32),\n", " array([0.43023413], dtype=float32),\n", " array([0.43097568], dtype=float32),\n", " array([0.4314532], dtype=float32),\n", " array([0.4314033], dtype=float32),\n", " array([0.43082014], dtype=float32),\n", " array([0.43014044], dtype=float32),\n", " array([0.42988652], dtype=float32),\n", " array([0.4301805], dtype=float32),\n", " array([0.4307109], dtype=float32),\n", " array([0.431059], dtype=float32),\n", " array([0.4310785], dtype=float32),\n", " array([0.43094414], dtype=float32),\n", " array([0.43082684], dtype=float32),\n", " array([0.43071404], dtype=float32),\n", " array([0.43060538], dtype=float32),\n", " array([0.43063235], dtype=float32),\n", " array([0.43091202], dtype=float32),\n", " array([0.4313794], dtype=float32),\n", " array([0.4316663], dtype=float32),\n", " array([0.431138], dtype=float32),\n", " array([0.42957577], dtype=float32),\n", " array([0.42777008], dtype=float32),\n", " array([0.42700928], dtype=float32),\n", " array([0.42789763], dtype=float32),\n", " array([0.42994413], dtype=float32),\n", " array([0.4321314], dtype=float32),\n", " array([0.43359584], dtype=float32),\n", " array([0.43386662], dtype=float32),\n", " array([0.4327987], dtype=float32),\n", " array([0.4306169], dtype=float32),\n", " array([0.42810342], dtype=float32),\n", " array([0.4264017], dtype=float32),\n", " array([0.42639768], dtype=float32),\n", " array([0.4281745], dtype=float32),\n", " array([0.4309964], dtype=float32),\n", " array([0.43368006], dtype=float32),\n", " array([0.43512177], dtype=float32),\n", " array([0.43475568], dtype=float32),\n", " array([0.43278524], dtype=float32),\n", " array([0.43012527], dtype=float32),\n", " array([0.42798704], dtype=float32),\n", " array([0.42726067], dtype=float32),\n", " array([0.42801246], dtype=float32),\n", " array([0.42955613], dtype=float32),\n", " array([0.43109512], dtype=float32),\n", " array([0.43220627], dtype=float32),\n", " array([0.43276286], dtype=float32),\n", " array([0.4327375], dtype=float32),\n", " array([0.43219763], dtype=float32),\n", " array([0.4313153], dtype=float32),\n", " array([0.4303774], dtype=float32),\n", " array([0.4297714], dtype=float32),\n", " array([0.42976746], dtype=float32),\n", " array([0.4302935], dtype=float32),\n", " array([0.4310517], dtype=float32),\n", " array([0.43178472], dtype=float32),\n", " array([0.4323415], dtype=float32),\n", " array([0.43261153], dtype=float32),\n", " array([0.4325141], dtype=float32),\n", " array([0.4320371], dtype=float32),\n", " array([0.4312907], dtype=float32),\n", " array([0.43053326], dtype=float32),\n", " array([0.4300949], dtype=float32),\n", " array([0.43018118], dtype=float32),\n", " array([0.4307371], dtype=float32),\n", " array([0.43149823], dtype=float32),\n", " array([0.4321439], dtype=float32),\n", " array([0.43233725], dtype=float32),\n", " array([0.43174583], dtype=float32),\n", " array([0.43027127], dtype=float32),\n", " array([0.42831066], dtype=float32),\n", " array([0.42666835], dtype=float32),\n", " array([0.42618218], dtype=float32),\n", " array([0.42728993], dtype=float32),\n", " array([0.4296954], dtype=float32),\n", " array([0.43238094], dtype=float32),\n", " array([0.43420702], dtype=float32),\n", " array([0.4345672], dtype=float32),\n", " array([0.43356144], dtype=float32),\n", " array([0.4318036], dtype=float32),\n", " array([0.43009862], dtype=float32),\n", " array([0.42902425], dtype=float32),\n", " array([0.42865673], dtype=float32),\n", " array([0.42878005], dtype=float32),\n", " array([0.42923227], dtype=float32),\n", " array([0.42986366], dtype=float32),\n", " array([0.43036836], dtype=float32),\n", " array([0.4304157], dtype=float32),\n", " array([0.4298911], dtype=float32),\n", " array([0.42902303], dtype=float32),\n", " array([0.42834735], dtype=float32),\n", " array([0.42839867], dtype=float32),\n", " array([0.4292522], dtype=float32),\n", " array([0.43038198], dtype=float32),\n", " array([0.43108362], dtype=float32),\n", " array([0.43098876], dtype=float32),\n", " array([0.4302291], dtype=float32),\n", " array([0.4293118], dtype=float32),\n", " array([0.42883798], dtype=float32),\n", " array([0.42917395], dtype=float32),\n", " array([0.43021914], dtype=float32),\n", " array([0.43148667], dtype=float32),\n", " array([0.43234053], dtype=float32),\n", " array([0.43221313], dtype=float32),\n", " array([0.43093786], dtype=float32),\n", " array([0.42901018], dtype=float32),\n", " array([0.42739916], dtype=float32),\n", " array([0.42684948], dtype=float32),\n", " array([0.42733884], dtype=float32),\n", " array([0.42835304], dtype=float32),\n", " array([0.42947292], dtype=float32),\n", " array([0.43050146], dtype=float32),\n", " array([0.431216], dtype=float32),\n", " array([0.43136427], dtype=float32),\n", " array([0.43092662], dtype=float32),\n", " array([0.43024135], dtype=float32),\n", " array([0.42982823], dtype=float32),\n", " array([0.4300614], dtype=float32),\n", " array([0.43090853], dtype=float32),\n", " array([0.4318901], dtype=float32),\n", " array([0.43240032], dtype=float32),\n", " array([0.43214568], dtype=float32),\n", " array([0.43132165], dtype=float32),\n", " array([0.43040362], dtype=float32),\n", " array([0.42972636], dtype=float32),\n", " array([0.42922434], dtype=float32),\n", " array([0.4285978], dtype=float32),\n", " array([0.42781046], dtype=float32),\n", " array([0.42729217], dtype=float32),\n", " array([0.42757997], dtype=float32),\n", " array([0.4287034], dtype=float32),\n", " array([0.4300258], dtype=float32),\n", " array([0.4308255], dtype=float32),\n", " array([0.43090695], dtype=float32),\n", " array([0.43063965], dtype=float32),\n", " array([0.4304987], dtype=float32),\n", " array([0.43054932], dtype=float32),\n", " array([0.4303976], dtype=float32),\n", " array([0.4296992], dtype=float32),\n", " array([0.4286403], dtype=float32),\n", " array([0.4278902], dtype=float32),\n", " array([0.42808503], dtype=float32),\n", " array([0.42926434], dtype=float32),\n", " array([0.43081152], dtype=float32),\n", " array([0.4318273], dtype=float32),\n", " array([0.43167987], dtype=float32),\n", " array([0.43043935], dtype=float32),\n", " array([0.42885232], dtype=float32),\n", " array([0.42790493], dtype=float32),\n", " array([0.42822215], dtype=float32),\n", " array([0.42973337], dtype=float32),\n", " array([0.43181035], dtype=float32),\n", " array([0.43357933], dtype=float32),\n", " array([0.434211], dtype=float32),\n", " array([0.4332293], dtype=float32),\n", " array([0.4308124], dtype=float32),\n", " array([0.4278903], dtype=float32),\n", " array([0.4258025], dtype=float32),\n", " array([0.42555436], dtype=float32),\n", " array([0.42715216], dtype=float32),\n", " array([0.42962307], dtype=float32),\n", " array([0.43166584], dtype=float32),\n", " array([0.43236524], dtype=float32),\n", " array([0.43152806], dtype=float32),\n", " array([0.4295788], dtype=float32),\n", " array([0.42724073], dtype=float32),\n", " array([0.42523327], dtype=float32),\n", " array([0.4241154], dtype=float32),\n", " array([0.42424434], dtype=float32),\n", " array([0.42569482], dtype=float32),\n", " array([0.42810863], dtype=float32),\n", " array([0.43074504], dtype=float32),\n", " array([0.43279183], dtype=float32),\n", " array([0.43372533], dtype=float32),\n", " array([0.43352154], dtype=float32),\n", " array([0.43259797], dtype=float32),\n", " array([0.4315558], dtype=float32),\n", " array([0.43089807], dtype=float32),\n", " array([0.43083456], dtype=float32),\n", " array([0.43120155], dtype=float32),\n", " array([0.43159732], dtype=float32),\n", " array([0.43166107], dtype=float32),\n", " array([0.43125802], dtype=float32),\n", " array([0.43048778], dtype=float32),\n", " array([0.42958617], dtype=float32),\n", " array([0.428821], dtype=float32),\n", " array([0.42841953], dtype=float32),\n", " array([0.42850748], dtype=float32),\n", " array([0.42903107], dtype=float32),\n", " array([0.42976415], dtype=float32),\n", " array([0.43043357], dtype=float32),\n", " array([0.43084416], dtype=float32),\n", " array([0.430951], dtype=float32),\n", " array([0.43083617], dtype=float32),\n", " array([0.43063766], dtype=float32),\n", " array([0.43050334], dtype=float32),\n", " array([0.43053076], dtype=float32),\n", " array([0.4307175], dtype=float32),\n", " array([0.43097898], dtype=float32),\n", " array([0.43121466], dtype=float32),\n", " array([0.43135616], dtype=float32),\n", " array([0.4313712], dtype=float32),\n", " array([0.4312545], dtype=float32),\n", " array([0.43102777], dtype=float32),\n", " array([0.43075517], dtype=float32),\n", " array([0.4305488], dtype=float32),\n", " array([0.430526], dtype=float32),\n", " array([0.43072462], dtype=float32),\n", " array([0.43104604], dtype=float32),\n", " array([0.43129563], dtype=float32),\n", " array([0.4313045], dtype=float32),\n", " array([0.431038], dtype=float32),\n", " array([0.43062612], dtype=float32),\n", " array([0.43029645], dtype=float32),\n", " array([0.43023807], dtype=float32),\n", " array([0.43050054], dtype=float32),\n", " array([0.43097997], dtype=float32),\n", " array([0.43145633], dtype=float32),\n", " array([0.43167284], dtype=float32),\n", " array([0.43141818], dtype=float32),\n", " array([0.43056494], dtype=float32),\n", " array([0.4291331], dtype=float32),\n", " array([0.42740452], dtype=float32),\n", " array([0.42595255], dtype=float32),\n", " array([0.42546013], dtype=float32),\n", " array([0.42636454], dtype=float32),\n", " array([0.4285342], dtype=float32),\n", " array([0.43125463], dtype=float32),\n", " array([0.4335604], dtype=float32),\n", " array([0.43467], dtype=float32),\n", " array([0.43428284], dtype=float32),\n", " array([0.43266085], dtype=float32),\n", " array([0.43051156], dtype=float32),\n", " array([0.42870954], dtype=float32),\n", " array([0.42793298], dtype=float32),\n", " array([0.42836732], dtype=float32),\n", " array([0.4296465], dtype=float32),\n", " array([0.4310863], dtype=float32),\n", " array([0.43204543], dtype=float32),\n", " array([0.43218625], dtype=float32),\n", " array([0.43153548], dtype=float32),\n", " array([0.43038574], dtype=float32),\n", " array([0.42915207], dtype=float32),\n", " array([0.4282593], dtype=float32),\n", " array([0.42803293], dtype=float32),\n", " array([0.4285797], dtype=float32),\n", " array([0.42972213], dtype=float32),\n", " array([0.43105927], dtype=float32),\n", " array([0.43213773], dtype=float32),\n", " array([0.43264067], dtype=float32),\n", " array([0.43248856], dtype=float32),\n", " array([0.43181828], dtype=float32),\n", " array([0.4308869], dtype=float32),\n", " array([0.42996088], dtype=float32),\n", " array([0.4292451], dtype=float32),\n", " array([0.42887634], dtype=float32),\n", " array([0.42894226], dtype=float32),\n", " array([0.42946315], dtype=float32),\n", " array([0.43033263], dtype=float32),\n", " array([0.43127868], dtype=float32),\n", " array([0.43191895], dtype=float32),\n", " array([0.43192247], dtype=float32),\n", " array([0.43118614], dtype=float32),\n", " array([0.4299216], dtype=float32),\n", " array([0.42860928], dtype=float32),\n", " array([0.4278087], dtype=float32),\n", " array([0.42788383], dtype=float32),\n", " array([0.4287987], dtype=float32),\n", " array([0.43010077], dtype=float32),\n", " array([0.431099], dtype=float32),\n", " array([0.43118295], dtype=float32),\n", " array([0.4301505], dtype=float32),\n", " array([0.4283649], dtype=float32),\n", " array([0.4266468], dtype=float32),\n", " array([0.42590934], dtype=float32),\n", " array([0.4266677], dtype=float32),\n", " ...],\n", " 'mean_module.constant': [array(0.0200181, dtype=float32),\n", " array(-0.03967667, dtype=float32),\n", " array(-0.08723032, dtype=float32),\n", " array(-0.10539389, dtype=float32),\n", " array(-0.10046721, dtype=float32),\n", " array(-0.08439076, dtype=float32),\n", " array(-0.06319046, dtype=float32),\n", " array(-0.04134703, dtype=float32),\n", " array(-0.02177656, dtype=float32),\n", " array(-0.00820231, dtype=float32),\n", " array(-0.00244975, dtype=float32),\n", " array(-0.00477457, dtype=float32),\n", " array(-0.01263547, dtype=float32),\n", " array(-0.02422035, dtype=float32),\n", " array(-0.03750122, dtype=float32),\n", " array(-0.04890847, dtype=float32),\n", " array(-0.05700445, dtype=float32),\n", " array(-0.05970538, dtype=float32),\n", " array(-0.05657649, dtype=float32),\n", " array(-0.04830217, dtype=float32),\n", " array(-0.03783429, dtype=float32),\n", " array(-0.0280273, dtype=float32),\n", " array(-0.01977634, dtype=float32),\n", " array(-0.01487792, dtype=float32),\n", " array(-0.01616299, dtype=float32),\n", " array(-0.02279222, dtype=float32),\n", " array(-0.03126097, dtype=float32),\n", " array(-0.04023921, dtype=float32),\n", " array(-0.04949093, dtype=float32),\n", " array(-0.05587292, dtype=float32),\n", " array(-0.05720174, dtype=float32),\n", " array(-0.05386651, dtype=float32),\n", " array(-0.04659116, dtype=float32),\n", " array(-0.03802872, dtype=float32),\n", " array(-0.03070295, dtype=float32),\n", " array(-0.02554882, dtype=float32),\n", " array(-0.02201617, dtype=float32),\n", " array(-0.02139008, dtype=float32),\n", " array(-0.02580369, dtype=float32),\n", " array(-0.03354883, dtype=float32),\n", " array(-0.04155934, dtype=float32),\n", " array(-0.04594409, dtype=float32),\n", " array(-0.04807878, dtype=float32),\n", " array(-0.04908609, dtype=float32),\n", " array(-0.04780614, dtype=float32),\n", " array(-0.04400325, dtype=float32),\n", " array(-0.0390259, dtype=float32),\n", " array(-0.03363872, dtype=float32),\n", " array(-0.03114963, dtype=float32),\n", " array(-0.03159857, dtype=float32),\n", " array(-0.03503942, dtype=float32),\n", " array(-0.03883171, dtype=float32),\n", " array(-0.0422647, dtype=float32),\n", " array(-0.04588461, dtype=float32),\n", " array(-0.04782128, dtype=float32),\n", " array(-0.04731369, dtype=float32),\n", " array(-0.04387164, dtype=float32),\n", " array(-0.03893006, dtype=float32),\n", " array(-0.03553104, dtype=float32),\n", " array(-0.03482735, dtype=float32),\n", " array(-0.03623068, dtype=float32),\n", " array(-0.03794611, dtype=float32),\n", " array(-0.03945363, dtype=float32),\n", " array(-0.04127395, dtype=float32),\n", " array(-0.04279339, dtype=float32),\n", " array(-0.04366434, dtype=float32),\n", " array(-0.04288626, dtype=float32),\n", " array(-0.03982782, dtype=float32),\n", " array(-0.03679621, dtype=float32),\n", " array(-0.03480375, dtype=float32),\n", " array(-0.0332489, dtype=float32),\n", " array(-0.03346705, dtype=float32),\n", " array(-0.03516233, dtype=float32),\n", " array(-0.03773737, dtype=float32),\n", " array(-0.03851902, dtype=float32),\n", " array(-0.03991544, dtype=float32),\n", " array(-0.04023826, dtype=float32),\n", " array(-0.04077399, dtype=float32),\n", " array(-0.04028416, dtype=float32),\n", " array(-0.03958869, dtype=float32),\n", " array(-0.03859901, dtype=float32),\n", " array(-0.03775191, dtype=float32),\n", " array(-0.03752589, dtype=float32),\n", " array(-0.03782856, dtype=float32),\n", " array(-0.03873527, dtype=float32),\n", " array(-0.03909385, dtype=float32),\n", " array(-0.03822815, dtype=float32),\n", " array(-0.0381465, dtype=float32),\n", " array(-0.03861248, dtype=float32),\n", " array(-0.03911841, dtype=float32),\n", " array(-0.03818607, dtype=float32),\n", " array(-0.03718007, dtype=float32),\n", " array(-0.03629613, dtype=float32),\n", " array(-0.0366286, dtype=float32),\n", " array(-0.03695202, dtype=float32),\n", " array(-0.03774226, dtype=float32),\n", " array(-0.03797603, dtype=float32),\n", " array(-0.03764701, dtype=float32),\n", " array(-0.0383029, dtype=float32),\n", " array(-0.0401479, dtype=float32),\n", " array(-0.04114389, dtype=float32),\n", " array(-0.04218912, dtype=float32),\n", " array(-0.04158187, dtype=float32),\n", " array(-0.03957188, dtype=float32),\n", " array(-0.0376544, dtype=float32),\n", " array(-0.03455389, dtype=float32),\n", " array(-0.03340864, dtype=float32),\n", " array(-0.03415501, dtype=float32),\n", " array(-0.03736556, dtype=float32),\n", " array(-0.04043138, dtype=float32),\n", " array(-0.04277933, dtype=float32),\n", " array(-0.0430336, dtype=float32),\n", " array(-0.04176891, dtype=float32),\n", " array(-0.03987885, dtype=float32),\n", " array(-0.03890991, dtype=float32),\n", " array(-0.0386101, dtype=float32),\n", " array(-0.03875566, dtype=float32),\n", " array(-0.03914702, dtype=float32),\n", " array(-0.03972518, dtype=float32),\n", " array(-0.04018295, dtype=float32),\n", " array(-0.03983784, dtype=float32),\n", " array(-0.03904843, dtype=float32),\n", " array(-0.03792739, dtype=float32),\n", " array(-0.03699327, dtype=float32),\n", " array(-0.03663409, dtype=float32),\n", " array(-0.03731072, dtype=float32),\n", " array(-0.03898346, dtype=float32),\n", " array(-0.04036558, dtype=float32),\n", " array(-0.04127443, dtype=float32),\n", " array(-0.04117572, dtype=float32),\n", " array(-0.04104829, dtype=float32),\n", " array(-0.04124773, dtype=float32),\n", " array(-0.04003334, dtype=float32),\n", " array(-0.03821516, dtype=float32),\n", " array(-0.03669941, dtype=float32),\n", " array(-0.03664088, dtype=float32),\n", " array(-0.03792155, dtype=float32),\n", " array(-0.03938603, dtype=float32),\n", " array(-0.0407635, dtype=float32),\n", " array(-0.04043317, dtype=float32),\n", " array(-0.03999543, dtype=float32),\n", " array(-0.03970182, dtype=float32),\n", " array(-0.03938067, dtype=float32),\n", " array(-0.03901565, dtype=float32),\n", " array(-0.03826094, dtype=float32),\n", " array(-0.03779054, dtype=float32),\n", " array(-0.03853822, dtype=float32),\n", " array(-0.03983665, dtype=float32),\n", " array(-0.04194844, dtype=float32),\n", " array(-0.04204905, dtype=float32),\n", " array(-0.04059362, dtype=float32),\n", " array(-0.04003882, dtype=float32),\n", " array(-0.03931963, dtype=float32),\n", " array(-0.03874874, dtype=float32),\n", " array(-0.03814387, dtype=float32),\n", " array(-0.03807175, dtype=float32),\n", " array(-0.04012668, dtype=float32),\n", " array(-0.04105318, dtype=float32),\n", " array(-0.0412358, dtype=float32),\n", " array(-0.04031932, dtype=float32),\n", " array(-0.03890991, dtype=float32),\n", " array(-0.0366118, dtype=float32),\n", " array(-0.03467381, dtype=float32),\n", " array(-0.03564596, dtype=float32),\n", " array(-0.03831661, dtype=float32),\n", " array(-0.04106843, dtype=float32),\n", " array(-0.04240167, dtype=float32),\n", " array(-0.04141319, dtype=float32),\n", " array(-0.03936696, dtype=float32),\n", " array(-0.03689492, dtype=float32),\n", " array(-0.03669226, dtype=float32),\n", " array(-0.03844953, dtype=float32),\n", " array(-0.03897858, dtype=float32),\n", " array(-0.03949475, dtype=float32),\n", " array(-0.0404247, dtype=float32),\n", " array(-0.04043901, dtype=float32),\n", " array(-0.03967249, dtype=float32),\n", " array(-0.03893995, dtype=float32),\n", " array(-0.03839076, dtype=float32),\n", " array(-0.03928721, dtype=float32),\n", " array(-0.04029024, dtype=float32),\n", " array(-0.04054606, dtype=float32),\n", " array(-0.03978062, dtype=float32),\n", " array(-0.03840435, dtype=float32),\n", " array(-0.03790939, dtype=float32),\n", " array(-0.03956008, dtype=float32),\n", " array(-0.04049575, dtype=float32),\n", " array(-0.04035378, dtype=float32),\n", " array(-0.03966641, dtype=float32),\n", " array(-0.03863215, dtype=float32),\n", " array(-0.03743625, dtype=float32),\n", " array(-0.03637064, dtype=float32),\n", " array(-0.03689456, dtype=float32),\n", " array(-0.03864539, dtype=float32),\n", " array(-0.04089367, dtype=float32),\n", " array(-0.04119051, dtype=float32),\n", " array(-0.03961718, dtype=float32),\n", " array(-0.03822923, dtype=float32),\n", " array(-0.0375849, dtype=float32),\n", " array(-0.03821528, dtype=float32),\n", " array(-0.03875577, dtype=float32),\n", " array(-0.03942227, dtype=float32),\n", " array(-0.04038513, dtype=float32),\n", " array(-0.04072607, dtype=float32),\n", " array(-0.04027414, dtype=float32),\n", " array(-0.03855753, dtype=float32),\n", " array(-0.03750002, dtype=float32),\n", " array(-0.03763211, dtype=float32),\n", " array(-0.03922391, dtype=float32),\n", " array(-0.04091036, dtype=float32),\n", " array(-0.04148304, dtype=float32),\n", " array(-0.04081976, dtype=float32),\n", " array(-0.03835404, dtype=float32),\n", " array(-0.03637373, dtype=float32),\n", " array(-0.03644395, dtype=float32),\n", " array(-0.03907669, dtype=float32),\n", " array(-0.04184687, dtype=float32),\n", " array(-0.04058337, dtype=float32),\n", " array(-0.03839779, dtype=float32),\n", " array(-0.0368365, dtype=float32),\n", " array(-0.03703725, dtype=float32),\n", " array(-0.03827345, dtype=float32),\n", " array(-0.03721082, dtype=float32),\n", " array(-0.03732729, dtype=float32),\n", " array(-0.03939664, dtype=float32),\n", " array(-0.04133654, dtype=float32),\n", " array(-0.0419246, dtype=float32),\n", " array(-0.03894949, dtype=float32),\n", " array(-0.03608036, dtype=float32),\n", " array(-0.03481627, dtype=float32),\n", " array(-0.03580832, dtype=float32),\n", " array(-0.03952754, dtype=float32),\n", " array(-0.04244006, dtype=float32),\n", " array(-0.04276204, dtype=float32),\n", " array(-0.04012132, dtype=float32),\n", " array(-0.03611553, dtype=float32),\n", " array(-0.03430915, dtype=float32),\n", " array(-0.03550851, dtype=float32),\n", " array(-0.03898704, dtype=float32),\n", " array(-0.04288495, dtype=float32),\n", " array(-0.04463828, dtype=float32),\n", " array(-0.04201341, dtype=float32),\n", " array(-0.03709745, dtype=float32),\n", " array(-0.03385806, dtype=float32),\n", " array(-0.03403091, dtype=float32),\n", " array(-0.03741658, dtype=float32),\n", " array(-0.04048908, dtype=float32),\n", " array(-0.04199505, dtype=float32),\n", " array(-0.04168403, dtype=float32),\n", " array(-0.03961647, dtype=float32),\n", " array(-0.0371803, dtype=float32),\n", " array(-0.03494096, dtype=float32),\n", " array(-0.03453541, dtype=float32),\n", " array(-0.0376718, dtype=float32),\n", " array(-0.04154229, dtype=float32),\n", " array(-0.04362142, dtype=float32),\n", " array(-0.04165423, dtype=float32),\n", " array(-0.03655887, dtype=float32),\n", " array(-0.0338161, dtype=float32),\n", " array(-0.03459418, dtype=float32),\n", " array(-0.03836918, dtype=float32),\n", " array(-0.04227817, dtype=float32),\n", " array(-0.04254913, dtype=float32),\n", " array(-0.0409137, dtype=float32),\n", " array(-0.03836262, dtype=float32),\n", " array(-0.03779411, dtype=float32),\n", " array(-0.038288, dtype=float32),\n", " array(-0.03858125, dtype=float32),\n", " array(-0.03916514, dtype=float32),\n", " array(-0.03973961, dtype=float32),\n", " array(-0.04061139, dtype=float32),\n", " array(-0.04057491, dtype=float32),\n", " array(-0.03928173, dtype=float32),\n", " array(-0.03712773, dtype=float32),\n", " array(-0.03676486, dtype=float32),\n", " array(-0.03789282, dtype=float32),\n", " array(-0.0392381, dtype=float32),\n", " array(-0.04021943, dtype=float32),\n", " array(-0.03956366, dtype=float32),\n", " array(-0.0388751, dtype=float32),\n", " array(-0.03871119, dtype=float32),\n", " array(-0.03773296, dtype=float32),\n", " array(-0.03742015, dtype=float32),\n", " array(-0.03766954, dtype=float32),\n", " array(-0.03913641, dtype=float32),\n", " array(-0.0405494, dtype=float32),\n", " array(-0.04068005, dtype=float32),\n", " array(-0.0395447, dtype=float32),\n", " array(-0.03854656, dtype=float32),\n", " array(-0.03970313, dtype=float32),\n", " array(-0.04111838, dtype=float32),\n", " array(-0.04107273, dtype=float32),\n", " array(-0.04021239, dtype=float32),\n", " array(-0.03945708, dtype=float32),\n", " array(-0.04135323, dtype=float32),\n", " array(-0.04177141, dtype=float32),\n", " array(-0.0405153, dtype=float32),\n", " array(-0.03819311, dtype=float32),\n", " array(-0.037238, dtype=float32),\n", " array(-0.03836739, dtype=float32),\n", " array(-0.04052293, dtype=float32),\n", " array(-0.04142559, dtype=float32),\n", " array(-0.0416764, dtype=float32),\n", " array(-0.04026055, dtype=float32),\n", " array(-0.03749573, dtype=float32),\n", " array(-0.03651536, dtype=float32),\n", " array(-0.03764963, dtype=float32),\n", " array(-0.03993106, dtype=float32),\n", " array(-0.04152441, dtype=float32),\n", " array(-0.04033291, dtype=float32),\n", " array(-0.03862762, dtype=float32),\n", " array(-0.03696883, dtype=float32),\n", " array(-0.03638816, dtype=float32),\n", " array(-0.03714645, dtype=float32),\n", " array(-0.03904223, dtype=float32),\n", " array(-0.04099309, dtype=float32),\n", " array(-0.04121959, dtype=float32),\n", " array(-0.0394187, dtype=float32),\n", " array(-0.03726661, dtype=float32),\n", " array(-0.03648937, dtype=float32),\n", " array(-0.03759551, dtype=float32),\n", " array(-0.0396769, dtype=float32),\n", " array(-0.04095435, dtype=float32),\n", " array(-0.0409193, dtype=float32),\n", " array(-0.0393914, dtype=float32),\n", " array(-0.03750086, dtype=float32),\n", " array(-0.0368011, dtype=float32),\n", " array(-0.03730941, dtype=float32),\n", " array(-0.03835034, dtype=float32),\n", " array(-0.03984904, dtype=float32),\n", " array(-0.0402602, dtype=float32),\n", " array(-0.03950608, dtype=float32),\n", " array(-0.03851414, dtype=float32),\n", " array(-0.0375464, dtype=float32),\n", " array(-0.03770995, dtype=float32),\n", " array(-0.03849733, dtype=float32),\n", " array(-0.03961027, dtype=float32),\n", " array(-0.04055488, dtype=float32),\n", " array(-0.04050624, dtype=float32),\n", " array(-0.03974557, dtype=float32),\n", " array(-0.03860569, dtype=float32),\n", " array(-0.03845263, dtype=float32),\n", " array(-0.03938758, dtype=float32),\n", " array(-0.03994298, dtype=float32),\n", " array(-0.03944695, dtype=float32),\n", " array(-0.03880703, dtype=float32),\n", " array(-0.03851902, dtype=float32),\n", " array(-0.03894055, dtype=float32),\n", " array(-0.03994572, dtype=float32),\n", " array(-0.04001975, dtype=float32),\n", " array(-0.03867018, dtype=float32),\n", " array(-0.03874397, dtype=float32),\n", " array(-0.03925419, dtype=float32),\n", " array(-0.03908622, dtype=float32),\n", " array(-0.03895128, dtype=float32),\n", " array(-0.03941071, dtype=float32),\n", " array(-0.03983128, dtype=float32),\n", " array(-0.040048, dtype=float32),\n", " array(-0.03878522, dtype=float32),\n", " array(-0.03743231, dtype=float32),\n", " array(-0.03762746, dtype=float32),\n", " array(-0.04008889, dtype=float32),\n", " array(-0.04195535, dtype=float32),\n", " array(-0.04063773, dtype=float32),\n", " array(-0.03772986, dtype=float32),\n", " array(-0.03669381, dtype=float32),\n", " array(-0.03788304, dtype=float32),\n", " array(-0.04035497, dtype=float32),\n", " array(-0.04190624, dtype=float32),\n", " array(-0.04105687, dtype=float32),\n", " array(-0.03858614, dtype=float32),\n", " array(-0.03664446, dtype=float32),\n", " array(-0.03695011, dtype=float32),\n", " array(-0.03963506, dtype=float32),\n", " array(-0.04244816, dtype=float32),\n", " array(-0.04273283, dtype=float32),\n", " array(-0.04046631, dtype=float32),\n", " array(-0.03747594, dtype=float32),\n", " array(-0.03587282, dtype=float32),\n", " array(-0.0377928, dtype=float32),\n", " array(-0.04122949, dtype=float32),\n", " array(-0.04257858, dtype=float32),\n", " array(-0.04069781, dtype=float32),\n", " array(-0.03825486, dtype=float32),\n", " array(-0.0373944, dtype=float32),\n", " array(-0.03833079, dtype=float32),\n", " array(-0.03973913, dtype=float32),\n", " array(-0.04051137, dtype=float32),\n", " array(-0.03914309, dtype=float32),\n", " array(-0.03851783, dtype=float32),\n", " array(-0.03924501, dtype=float32),\n", " array(-0.03994286, dtype=float32),\n", " array(-0.04029012, dtype=float32),\n", " array(-0.0409832, dtype=float32),\n", " array(-0.04008722, dtype=float32),\n", " array(-0.03870177, dtype=float32),\n", " array(-0.03839993, dtype=float32),\n", " array(-0.0397346, dtype=float32),\n", " array(-0.04034889, dtype=float32),\n", " array(-0.04004216, dtype=float32),\n", " array(-0.03885424, dtype=float32),\n", " array(-0.03757584, dtype=float32),\n", " array(-0.03785169, dtype=float32),\n", " array(-0.04097772, dtype=float32),\n", " array(-0.04218912, dtype=float32),\n", " array(-0.04020274, dtype=float32),\n", " array(-0.03757429, dtype=float32),\n", " array(-0.03592396, dtype=float32),\n", " array(-0.03674138, dtype=float32),\n", " array(-0.04030085, dtype=float32),\n", " array(-0.04282212, dtype=float32),\n", " array(-0.04196143, dtype=float32),\n", " array(-0.03858006, dtype=float32),\n", " array(-0.03487325, dtype=float32),\n", " array(-0.03472984, dtype=float32),\n", " array(-0.03875756, dtype=float32),\n", " array(-0.043239, dtype=float32),\n", " array(-0.04420078, dtype=float32),\n", " array(-0.04106367, dtype=float32),\n", " array(-0.03628969, dtype=float32),\n", " array(-0.03444624, dtype=float32),\n", " array(-0.03701913, dtype=float32),\n", " array(-0.04182982, dtype=float32),\n", " array(-0.04455829, dtype=float32),\n", " array(-0.04333305, dtype=float32),\n", " array(-0.03954268, dtype=float32),\n", " array(-0.03617477, dtype=float32),\n", " array(-0.03567374, dtype=float32),\n", " array(-0.03854895, dtype=float32),\n", " array(-0.04314649, dtype=float32),\n", " array(-0.04540169, dtype=float32),\n", " array(-0.04273927, dtype=float32),\n", " array(-0.03759158, dtype=float32),\n", " array(-0.03498864, dtype=float32),\n", " array(-0.03684139, dtype=float32),\n", " array(-0.04162264, dtype=float32),\n", " array(-0.04485941, dtype=float32),\n", " array(-0.04342854, dtype=float32),\n", " array(-0.03872705, dtype=float32),\n", " array(-0.03483653, dtype=float32),\n", " array(-0.03504574, dtype=float32),\n", " array(-0.03836942, dtype=float32),\n", " array(-0.04141212, dtype=float32),\n", " array(-0.04209769, dtype=float32),\n", " array(-0.04099667, dtype=float32),\n", " array(-0.03818202, dtype=float32),\n", " array(-0.03554714, dtype=float32),\n", " array(-0.03557527, dtype=float32),\n", " array(-0.03818464, dtype=float32),\n", " array(-0.04158306, dtype=float32),\n", " array(-0.04399073, dtype=float32),\n", " array(-0.04352868, dtype=float32),\n", " array(-0.03933251, dtype=float32),\n", " array(-0.03504992, dtype=float32),\n", " array(-0.0348779, dtype=float32),\n", " array(-0.03883255, dtype=float32),\n", " array(-0.04327667, dtype=float32),\n", " array(-0.04465067, dtype=float32),\n", " array(-0.04154444, dtype=float32),\n", " array(-0.03707469, dtype=float32),\n", " array(-0.03574097, dtype=float32),\n", " array(-0.03866851, dtype=float32),\n", " array(-0.04215455, dtype=float32),\n", " array(-0.04263282, dtype=float32),\n", " array(-0.03989697, dtype=float32),\n", " array(-0.03668523, dtype=float32),\n", " array(-0.03583658, dtype=float32),\n", " array(-0.03865016, dtype=float32),\n", " array(-0.04232013, dtype=float32),\n", " array(-0.04301119, dtype=float32),\n", " array(-0.0405252, dtype=float32),\n", " array(-0.03719401, dtype=float32),\n", " array(-0.03612685, dtype=float32),\n", " array(-0.03820443, dtype=float32),\n", " array(-0.04102147, dtype=float32),\n", " array(-0.04200423, dtype=float32),\n", " array(-0.04067397, dtype=float32),\n", " array(-0.0382725, dtype=float32),\n", " array(-0.03726423, dtype=float32),\n", " array(-0.03876233, dtype=float32),\n", " array(-0.04034984, dtype=float32),\n", " array(-0.04052162, dtype=float32),\n", " array(-0.03968, dtype=float32),\n", " array(-0.03801942, dtype=float32),\n", " array(-0.03710496, dtype=float32),\n", " array(-0.03822815, dtype=float32),\n", " array(-0.04006839, dtype=float32),\n", " array(-0.04076266, dtype=float32),\n", " array(-0.04045105, dtype=float32),\n", " array(-0.03914249, dtype=float32),\n", " array(-0.03732777, dtype=float32),\n", " array(-0.0371573, dtype=float32),\n", " array(-0.03880119, dtype=float32),\n", " array(-0.0401051, dtype=float32),\n", " array(-0.03974307, dtype=float32),\n", " array(-0.03846192, dtype=float32),\n", " array(-0.03784716, dtype=float32),\n", " array(-0.03869689, dtype=float32),\n", " array(-0.04003036, dtype=float32),\n", " array(-0.04033148, dtype=float32),\n", " array(-0.03929722, dtype=float32),\n", " array(-0.03806901, dtype=float32),\n", " array(-0.03806674, dtype=float32),\n", " array(-0.03942585, dtype=float32),\n", " array(-0.04046977, dtype=float32),\n", " array(-0.03990901, dtype=float32),\n", " array(-0.03889155, dtype=float32),\n", " array(-0.03891659, dtype=float32),\n", " array(-0.04006779, dtype=float32),\n", " array(-0.04081726, dtype=float32),\n", " array(-0.03980303, dtype=float32),\n", " array(-0.03802931, dtype=float32),\n", " array(-0.03731787, dtype=float32),\n", " array(-0.03774703, dtype=float32),\n", " array(-0.03891861, dtype=float32),\n", " array(-0.04054534, dtype=float32),\n", " array(-0.04119289, dtype=float32),\n", " array(-0.03985965, dtype=float32),\n", " array(-0.037925, dtype=float32),\n", " array(-0.03664505, dtype=float32),\n", " array(-0.03639364, dtype=float32),\n", " array(-0.03803968, dtype=float32),\n", " array(-0.0409193, dtype=float32),\n", " array(-0.04233074, dtype=float32),\n", " array(-0.04048574, dtype=float32),\n", " array(-0.03729868, dtype=float32),\n", " array(-0.03580153, dtype=float32),\n", " array(-0.03708839, dtype=float32),\n", " array(-0.03988552, dtype=float32),\n", " array(-0.0416398, dtype=float32),\n", " array(-0.04069448, dtype=float32),\n", " array(-0.0380882, dtype=float32),\n", " array(-0.03653669, dtype=float32),\n", " array(-0.03774011, dtype=float32),\n", " array(-0.04060698, dtype=float32),\n", " array(-0.04199314, dtype=float32),\n", " array(-0.04041696, dtype=float32),\n", " array(-0.0376389, dtype=float32),\n", " array(-0.036502, dtype=float32),\n", " array(-0.03778899, dtype=float32),\n", " array(-0.0399605, dtype=float32),\n", " array(-0.04075193, dtype=float32),\n", " array(-0.0394547, dtype=float32),\n", " array(-0.03767884, dtype=float32),\n", " array(-0.03762805, dtype=float32),\n", " array(-0.03938568, dtype=float32),\n", " array(-0.04091156, dtype=float32),\n", " array(-0.04053843, dtype=float32),\n", " array(-0.03877068, dtype=float32),\n", " array(-0.03742468, dtype=float32),\n", " array(-0.03747225, dtype=float32),\n", " array(-0.03903782, dtype=float32),\n", " array(-0.04069066, dtype=float32),\n", " array(-0.0407815, dtype=float32),\n", " array(-0.03947341, dtype=float32),\n", " array(-0.03851533, dtype=float32),\n", " array(-0.03871953, dtype=float32),\n", " array(-0.03902495, dtype=float32),\n", " array(-0.0390414, dtype=float32),\n", " array(-0.03923166, dtype=float32),\n", " array(-0.03926873, dtype=float32),\n", " array(-0.03879035, dtype=float32),\n", " array(-0.0382942, dtype=float32),\n", " array(-0.03803289, dtype=float32),\n", " array(-0.03763044, dtype=float32),\n", " array(-0.0373137, dtype=float32),\n", " array(-0.03828907, dtype=float32),\n", " array(-0.04014218, dtype=float32),\n", " array(-0.04078043, dtype=float32),\n", " array(-0.03923905, dtype=float32),\n", " array(-0.03658295, dtype=float32),\n", " array(-0.03554928, dtype=float32),\n", " array(-0.03756845, dtype=float32),\n", " array(-0.04047012, dtype=float32),\n", " array(-0.0413332, dtype=float32),\n", " array(-0.03931224, dtype=float32),\n", " array(-0.03657389, dtype=float32),\n", " array(-0.03604746, dtype=float32),\n", " array(-0.03849363, dtype=float32),\n", " array(-0.04161537, dtype=float32),\n", " array(-0.04165792, dtype=float32),\n", " array(-0.03853333, dtype=float32),\n", " array(-0.03593624, dtype=float32),\n", " array(-0.03659821, dtype=float32),\n", " array(-0.03962219, dtype=float32),\n", " array(-0.04175889, dtype=float32),\n", " array(-0.04095745, dtype=float32),\n", " array(-0.03843904, dtype=float32),\n", " array(-0.03690958, dtype=float32),\n", " array(-0.03752553, dtype=float32),\n", " array(-0.0394783, dtype=float32),\n", " array(-0.04071856, dtype=float32),\n", " array(-0.04021633, dtype=float32),\n", " array(-0.03917575, dtype=float32),\n", " array(-0.03897989, dtype=float32),\n", " array(-0.0395087, dtype=float32),\n", " array(-0.03929031, dtype=float32),\n", " array(-0.0377847, dtype=float32),\n", " array(-0.03667748, dtype=float32),\n", " array(-0.03724623, dtype=float32),\n", " array(-0.03866291, dtype=float32),\n", " array(-0.03972602, dtype=float32),\n", " array(-0.03980315, dtype=float32),\n", " array(-0.03890932, dtype=float32),\n", " array(-0.03758264, dtype=float32),\n", " array(-0.03714597, dtype=float32),\n", " array(-0.03841186, dtype=float32),\n", " array(-0.03998876, dtype=float32),\n", " array(-0.04016459, dtype=float32),\n", " array(-0.03920376, dtype=float32),\n", " array(-0.03838444, dtype=float32),\n", " array(-0.0382576, dtype=float32),\n", " array(-0.03862369, dtype=float32),\n", " array(-0.03917754, dtype=float32),\n", " array(-0.03965783, dtype=float32),\n", " array(-0.03979707, dtype=float32),\n", " array(-0.039482, dtype=float32),\n", " array(-0.0389868, dtype=float32),\n", " array(-0.03896725, dtype=float32),\n", " array(-0.03945744, dtype=float32),\n", " array(-0.03914106, dtype=float32),\n", " array(-0.03752089, dtype=float32),\n", " array(-0.03660667, dtype=float32),\n", " array(-0.03812528, dtype=float32),\n", " array(-0.04058433, dtype=float32),\n", " array(-0.04131317, dtype=float32),\n", " array(-0.04013228, dtype=float32),\n", " array(-0.03852797, dtype=float32),\n", " array(-0.03738356, dtype=float32),\n", " array(-0.0375669, dtype=float32),\n", " array(-0.03937972, dtype=float32),\n", " array(-0.04106498, dtype=float32),\n", " array(-0.0405885, dtype=float32),\n", " array(-0.03841662, dtype=float32),\n", " array(-0.03714848, dtype=float32),\n", " array(-0.03829288, dtype=float32),\n", " array(-0.0404855, dtype=float32),\n", " array(-0.04127824, dtype=float32),\n", " array(-0.03978407, dtype=float32),\n", " array(-0.03745174, dtype=float32),\n", " array(-0.03661728, dtype=float32),\n", " array(-0.03826964, dtype=float32),\n", " array(-0.04100323, dtype=float32),\n", " array(-0.0422194, dtype=float32),\n", " array(-0.04088664, dtype=float32),\n", " array(-0.03868997, dtype=float32),\n", " array(-0.0378927, dtype=float32),\n", " array(-0.03879833, dtype=float32),\n", " array(-0.03989077, dtype=float32),\n", " array(-0.03986084, dtype=float32),\n", " array(-0.03900254, dtype=float32),\n", " array(-0.03875208, dtype=float32),\n", " array(-0.0399065, dtype=float32),\n", " array(-0.04138625, dtype=float32),\n", " array(-0.04130602, dtype=float32),\n", " array(-0.03941202, dtype=float32),\n", " array(-0.03753877, dtype=float32),\n", " array(-0.03750563, dtype=float32),\n", " array(-0.03919649, dtype=float32),\n", " array(-0.04078639, dtype=float32),\n", " array(-0.04063082, dtype=float32),\n", " array(-0.03894544, dtype=float32),\n", " array(-0.03760731, dtype=float32),\n", " array(-0.03812325, dtype=float32),\n", " array(-0.03987145, dtype=float32),\n", " array(-0.04077828, dtype=float32),\n", " array(-0.03976417, dtype=float32),\n", " array(-0.03796446, dtype=float32),\n", " array(-0.0373013, dtype=float32),\n", " array(-0.03838718, dtype=float32),\n", " array(-0.04002643, dtype=float32),\n", " array(-0.04065669, dtype=float32),\n", " array(-0.04000258, dtype=float32),\n", " array(-0.03882515, dtype=float32),\n", " array(-0.03800762, dtype=float32),\n", " array(-0.0381825, dtype=float32),\n", " array(-0.03907454, dtype=float32),\n", " array(-0.03952909, dtype=float32),\n", " array(-0.03894746, dtype=float32),\n", " array(-0.03821993, dtype=float32),\n", " array(-0.03857934, dtype=float32),\n", " array(-0.03962469, dtype=float32),\n", " array(-0.03979647, dtype=float32),\n", " array(-0.03885865, dtype=float32),\n", " array(-0.03815389, dtype=float32),\n", " array(-0.0384655, dtype=float32),\n", " array(-0.0391022, dtype=float32),\n", " array(-0.03913832, dtype=float32),\n", " array(-0.03869534, dtype=float32),\n", " array(-0.03851497, dtype=float32),\n", " array(-0.03886092, dtype=float32),\n", " array(-0.03937161, dtype=float32),\n", " array(-0.03950512, dtype=float32),\n", " array(-0.03908622, dtype=float32),\n", " array(-0.03851283, dtype=float32),\n", " array(-0.03831244, dtype=float32),\n", " array(-0.03869581, dtype=float32),\n", " array(-0.03935254, dtype=float32),\n", " array(-0.03977072, dtype=float32),\n", " array(-0.0397085, dtype=float32),\n", " array(-0.03934252, dtype=float32),\n", " array(-0.03920233, dtype=float32),\n", " array(-0.03947294, dtype=float32),\n", " array(-0.03966534, dtype=float32),\n", " array(-0.03936899, dtype=float32),\n", " array(-0.0388068, dtype=float32),\n", " array(-0.03861964, dtype=float32),\n", " array(-0.03908873, dtype=float32),\n", " array(-0.03974867, dtype=float32),\n", " array(-0.03994799, dtype=float32),\n", " array(-0.03955674, dtype=float32),\n", " array(-0.03925526, dtype=float32),\n", " array(-0.03959525, dtype=float32),\n", " array(-0.03997695, dtype=float32),\n", " array(-0.03960955, dtype=float32),\n", " array(-0.03864181, dtype=float32),\n", " array(-0.03759933, dtype=float32),\n", " array(-0.03691983, dtype=float32),\n", " array(-0.03740501, dtype=float32),\n", " array(-0.03946424, dtype=float32),\n", " array(-0.04157901, dtype=float32),\n", " array(-0.0414418, dtype=float32),\n", " array(-0.03908038, dtype=float32),\n", " array(-0.03713441, dtype=float32),\n", " array(-0.03736901, dtype=float32),\n", " array(-0.03866887, dtype=float32),\n", " array(-0.03940177, dtype=float32),\n", " array(-0.03952003, dtype=float32),\n", " array(-0.03956115, dtype=float32),\n", " array(-0.03950632, dtype=float32),\n", " array(-0.03911889, dtype=float32),\n", " array(-0.03865755, dtype=float32),\n", " array(-0.03888774, dtype=float32),\n", " array(-0.0398401, dtype=float32),\n", " array(-0.04030275, dtype=float32),\n", " array(-0.03940701, dtype=float32),\n", " array(-0.03800976, dtype=float32),\n", " array(-0.03763545, dtype=float32),\n", " array(-0.03853524, dtype=float32),\n", " array(-0.03945601, dtype=float32),\n", " array(-0.0393976, dtype=float32),\n", " array(-0.03878689, dtype=float32),\n", " array(-0.03851521, dtype=float32),\n", " array(-0.03859413, dtype=float32),\n", " array(-0.03846169, dtype=float32),\n", " array(-0.03800106, dtype=float32),\n", " array(-0.03780186, dtype=float32),\n", " array(-0.03841734, dtype=float32),\n", " array(-0.03943014, dtype=float32),\n", " array(-0.03978801, dtype=float32),\n", " array(-0.03930807, dtype=float32),\n", " array(-0.03890753, dtype=float32),\n", " array(-0.03907025, dtype=float32),\n", " array(-0.03905523, dtype=float32),\n", " array(-0.03835237, dtype=float32),\n", " array(-0.03796601, dtype=float32),\n", " array(-0.03890789, dtype=float32),\n", " array(-0.04030418, dtype=float32),\n", " array(-0.0405333, dtype=float32),\n", " array(-0.03946424, dtype=float32),\n", " array(-0.0384388, dtype=float32),\n", " array(-0.0383091, dtype=float32),\n", " array(-0.03860784, dtype=float32),\n", " array(-0.03879559, dtype=float32),\n", " array(-0.03905559, dtype=float32),\n", " array(-0.0394212, dtype=float32),\n", " array(-0.03937495, dtype=float32),\n", " array(-0.03867781, dtype=float32),\n", " array(-0.03793287, dtype=float32),\n", " array(-0.03792441, dtype=float32),\n", " array(-0.0385766, dtype=float32),\n", " array(-0.0390929, dtype=float32),\n", " array(-0.03912044, dtype=float32),\n", " array(-0.03899407, dtype=float32),\n", " array(-0.0388571, dtype=float32),\n", " array(-0.03855085, dtype=float32),\n", " array(-0.03832221, dtype=float32),\n", " array(-0.03869116, dtype=float32),\n", " array(-0.03948474, dtype=float32),\n", " array(-0.03980887, dtype=float32),\n", " array(-0.03921211, dtype=float32),\n", " array(-0.03840375, dtype=float32),\n", " array(-0.03834105, dtype=float32),\n", " array(-0.03896022, dtype=float32),\n", " array(-0.03947377, dtype=float32),\n", " array(-0.03953993, dtype=float32),\n", " array(-0.03936386, dtype=float32),\n", " array(-0.03909266, dtype=float32),\n", " array(-0.03884292, dtype=float32),\n", " array(-0.03886735, dtype=float32),\n", " array(-0.03918612, dtype=float32),\n", " array(-0.03942156, dtype=float32),\n", " array(-0.03933465, dtype=float32),\n", " array(-0.03925633, dtype=float32),\n", " array(-0.03953314, dtype=float32),\n", " array(-0.03977251, dtype=float32),\n", " array(-0.03940701, dtype=float32),\n", " array(-0.038849, dtype=float32),\n", " array(-0.03901553, dtype=float32),\n", " array(-0.03982854, dtype=float32),\n", " array(-0.04014111, dtype=float32),\n", " array(-0.03927684, dtype=float32),\n", " array(-0.03803921, dtype=float32),\n", " array(-0.03786016, dtype=float32),\n", " array(-0.03899121, dtype=float32),\n", " array(-0.04002571, dtype=float32),\n", " array(-0.03962982, dtype=float32),\n", " array(-0.03833544, dtype=float32),\n", " array(-0.03779697, dtype=float32),\n", " array(-0.03841448, dtype=float32),\n", " array(-0.03886378, dtype=float32),\n", " array(-0.03826559, dtype=float32),\n", " array(-0.03764451, dtype=float32),\n", " array(-0.03839517, dtype=float32),\n", " array(-0.04008543, dtype=float32),\n", " array(-0.04069912, dtype=float32),\n", " array(-0.03919363, dtype=float32),\n", " array(-0.0370791, dtype=float32),\n", " array(-0.0367378, dtype=float32),\n", " array(-0.03853512, dtype=float32),\n", " array(-0.04045653, dtype=float32),\n", " array(-0.04058599, dtype=float32),\n", " array(-0.03933215, dtype=float32),\n", " array(-0.03871369, dtype=float32),\n", " array(-0.03949046, dtype=float32),\n", " array(-0.04015362, dtype=float32),\n", " array(-0.03932142, dtype=float32),\n", " array(-0.03795981, dtype=float32),\n", " array(-0.0380789, dtype=float32),\n", " array(-0.03978729, dtype=float32),\n", " array(-0.04090893, dtype=float32),\n", " array(-0.03981602, dtype=float32),\n", " array(-0.03773379, dtype=float32),\n", " array(-0.03721571, dtype=float32),\n", " array(-0.03876197, dtype=float32),\n", " array(-0.04024124, dtype=float32),\n", " array(-0.03983569, dtype=float32),\n", " array(-0.03837359, dtype=float32),\n", " array(-0.03800356, dtype=float32),\n", " array(-0.03921592, dtype=float32),\n", " array(-0.04030764, dtype=float32),\n", " array(-0.03981924, dtype=float32),\n", " array(-0.03840399, dtype=float32),\n", " array(-0.03772807, dtype=float32),\n", " array(-0.0383029, dtype=float32),\n", " array(-0.03916407, dtype=float32),\n", " array(-0.0393064, dtype=float32),\n", " array(-0.03875256, dtype=float32),\n", " array(-0.03828001, dtype=float32),\n", " array(-0.03854132, dtype=float32),\n", " array(-0.03938913, dtype=float32),\n", " array(-0.03991497, dtype=float32),\n", " array(-0.03934336, dtype=float32),\n", " array(-0.03802013, dtype=float32),\n", " array(-0.03729761, dtype=float32),\n", " array(-0.03806472, dtype=float32),\n", " array(-0.03952181, dtype=float32),\n", " array(-0.04000449, dtype=float32),\n", " array(-0.03907394, dtype=float32),\n", " array(-0.03807342, dtype=float32),\n", " array(-0.03832817, dtype=float32),\n", " array(-0.03935003, dtype=float32),\n", " array(-0.03954494, dtype=float32),\n", " array(-0.03848588, dtype=float32),\n", " array(-0.03760314, dtype=float32),\n", " array(-0.0382762, dtype=float32),\n", " array(-0.03981268, dtype=float32),\n", " array(-0.04023063, dtype=float32),\n", " array(-0.03888977, dtype=float32),\n", " array(-0.03741741, dtype=float32),\n", " array(-0.03769171, dtype=float32),\n", " array(-0.03940511, dtype=float32),\n", " array(-0.04053247, dtype=float32),\n", " array(-0.03998542, dtype=float32),\n", " array(-0.03883183, dtype=float32),\n", " array(-0.03860521, dtype=float32),\n", " array(-0.03919733, dtype=float32),\n", " array(-0.0392698, dtype=float32),\n", " array(-0.03836918, dtype=float32),\n", " array(-0.03776062, dtype=float32),\n", " array(-0.03873563, dtype=float32),\n", " array(-0.0405978, dtype=float32),\n", " array(-0.04128397, dtype=float32),\n", " array(-0.04005647, dtype=float32),\n", " array(-0.03863096, dtype=float32),\n", " array(-0.03894842, dtype=float32),\n", " array(-0.04044509, dtype=float32),\n", " array(-0.04070854, dtype=float32),\n", " array(-0.03890598, dtype=float32),\n", " array(-0.03733814, dtype=float32),\n", " array(-0.03845358, dtype=float32),\n", " array(-0.04106784, dtype=float32),\n", " array(-0.04165363, dtype=float32),\n", " array(-0.03938961, dtype=float32),\n", " array(-0.03751326, dtype=float32),\n", " array(-0.03863895, dtype=float32),\n", " array(-0.0409174, dtype=float32),\n", " array(-0.0407964, dtype=float32),\n", " array(-0.03822815, dtype=float32),\n", " array(-0.03677368, dtype=float32),\n", " array(-0.03849626, dtype=float32),\n", " array(-0.04101992, dtype=float32),\n", " array(-0.04096961, dtype=float32),\n", " array(-0.03867602, dtype=float32),\n", " array(-0.03742695, dtype=float32),\n", " array(-0.03867471, dtype=float32),\n", " array(-0.04023325, dtype=float32),\n", " array(-0.03972137, dtype=float32),\n", " array(-0.03792894, dtype=float32),\n", " array(-0.03736794, dtype=float32),\n", " array(-0.03857172, dtype=float32),\n", " array(-0.03964245, dtype=float32),\n", " array(-0.03924644, dtype=float32),\n", " array(-0.0383383, dtype=float32),\n", " array(-0.03838074, dtype=float32),\n", " array(-0.03912711, dtype=float32),\n", " array(-0.03922164, dtype=float32),\n", " array(-0.03839052, dtype=float32),\n", " array(-0.03784001, dtype=float32),\n", " array(-0.03844988, dtype=float32),\n", " array(-0.03953087, dtype=float32),\n", " array(-0.03984773, dtype=float32),\n", " array(-0.0392698, dtype=float32),\n", " array(-0.03871095, dtype=float32),\n", " array(-0.03869474, dtype=float32),\n", " array(-0.03873575, dtype=float32),\n", " array(-0.0383184, dtype=float32),\n", " array(-0.03782845, dtype=float32),\n", " array(-0.03799832, dtype=float32),\n", " array(-0.0387882, dtype=float32),\n", " array(-0.0393846, dtype=float32),\n", " array(-0.03929543, dtype=float32),\n", " array(-0.03895831, dtype=float32),\n", " array(-0.03902197, dtype=float32),\n", " array(-0.0394063, dtype=float32),\n", " array(-0.03950024, dtype=float32),\n", " array(-0.0391041, dtype=float32),\n", " array(-0.0386914, dtype=float32),\n", " array(-0.0386703, dtype=float32),\n", " array(-0.03882802, dtype=float32),\n", " array(-0.03875124, dtype=float32),\n", " array(-0.03850567, dtype=float32),\n", " array(-0.03853178, dtype=float32),\n", " array(-0.03895628, dtype=float32),\n", " array(-0.03936422, dtype=float32),\n", " array(-0.03937006, dtype=float32),\n", " array(-0.03913307, dtype=float32),\n", " array(-0.03905249, dtype=float32),\n", " array(-0.03912079, dtype=float32),\n", " array(-0.03896415, dtype=float32),\n", " array(-0.03851664, dtype=float32),\n", " array(-0.03825366, dtype=float32),\n", " array(-0.03857601, dtype=float32),\n", " array(-0.03922749, dtype=float32),\n", " array(-0.03961885, dtype=float32),\n", " array(-0.03960049, dtype=float32),\n", " array(-0.03957176, dtype=float32),\n", " array(-0.03979504, dtype=float32),\n", " array(-0.03996158, dtype=float32),\n", " array(-0.03968537, dtype=float32),\n", " array(-0.039168, dtype=float32),\n", " array(-0.03897572, dtype=float32),\n", " array(-0.03925645, dtype=float32),\n", " array(-0.03957427, dtype=float32),\n", " array(-0.03954685, dtype=float32),\n", " array(-0.03932762, dtype=float32),\n", " array(-0.03928065, dtype=float32),\n", " array(-0.0394156, dtype=float32),\n", " array(-0.03944564, dtype=float32),\n", " array(-0.03931236, dtype=float32),\n", " array(-0.03930068, dtype=float32),\n", " array(-0.03956902, dtype=float32),\n", " array(-0.03986335, dtype=float32),\n", " array(-0.03985441, dtype=float32),\n", " array(-0.03956866, dtype=float32),\n", " array(-0.03927088, dtype=float32),\n", " array(-0.03903377, dtype=float32),\n", " array(-0.03871143, dtype=float32),\n", " array(-0.03834701, dtype=float32),\n", " array(-0.03826797, dtype=float32),\n", " array(-0.03861678, dtype=float32),\n", " array(-0.03905404, dtype=float32),\n", " array(-0.03915489, dtype=float32),\n", " array(-0.03895104, dtype=float32),\n", " array(-0.03883004, dtype=float32),\n", " array(-0.03893912, dtype=float32),\n", " array(-0.03896487, dtype=float32),\n", " array(-0.03859019, dtype=float32),\n", " array(-0.03797483, dtype=float32),\n", " array(-0.03760529, dtype=float32),\n", " array(-0.03774047, dtype=float32),\n", " array(-0.03820169, dtype=float32),\n", " array(-0.03864706, dtype=float32),\n", " array(-0.03885806, dtype=float32),\n", " array(-0.03880143, dtype=float32),\n", " array(-0.03860569, dtype=float32),\n", " array(-0.03849089, dtype=float32),\n", " array(-0.03858817, dtype=float32),\n", " array(-0.03878701, dtype=float32),\n", " array(-0.03885329, dtype=float32),\n", " ...],\n", " 'covar_module.mixture_weights': [array([0.7347616], dtype=float32),\n", " array([0.6839323], dtype=float32),\n", " array([0.63559526], dtype=float32),\n", " array([0.58985114], dtype=float32),\n", " array([0.54691046], dtype=float32),\n", " array([0.5068499], dtype=float32),\n", " array([0.46949553], dtype=float32),\n", " array([0.4343483], dtype=float32),\n", " array([0.40149093], dtype=float32),\n", " array([0.37086368], dtype=float32),\n", " array([0.34259918], dtype=float32),\n", " array([0.3162862], dtype=float32),\n", " array([0.2917886], dtype=float32),\n", " array([0.26909676], dtype=float32),\n", " array([0.24812995], dtype=float32),\n", " array([0.22883604], dtype=float32),\n", " array([0.21104366], dtype=float32),\n", " array([0.19461492], dtype=float32),\n", " array([0.17953289], dtype=float32),\n", " array([0.16560112], dtype=float32),\n", " array([0.15284233], dtype=float32),\n", " array([0.14122175], dtype=float32),\n", " array([0.13055222], dtype=float32),\n", " array([0.12094312], dtype=float32),\n", " array([0.11213627], dtype=float32),\n", " array([0.1040785], dtype=float32),\n", " array([0.09671624], dtype=float32),\n", " array([0.08993911], dtype=float32),\n", " array([0.08375356], dtype=float32),\n", " array([0.0780656], dtype=float32),\n", " array([0.07277866], dtype=float32),\n", " array([0.06788274], dtype=float32),\n", " array([0.06339081], dtype=float32),\n", " array([0.05929431], dtype=float32),\n", " array([0.05551264], dtype=float32),\n", " array([0.0520482], dtype=float32),\n", " array([0.0488765], dtype=float32),\n", " array([0.04601723], dtype=float32),\n", " array([0.04342069], dtype=float32),\n", " array([0.04101533], dtype=float32),\n", " array([0.03877807], dtype=float32),\n", " array([0.03672723], dtype=float32),\n", " array([0.03484517], dtype=float32),\n", " array([0.0331106], dtype=float32),\n", " array([0.03149256], dtype=float32),\n", " array([0.02997206], dtype=float32),\n", " array([0.02854297], dtype=float32),\n", " array([0.0271967], dtype=float32),\n", " array([0.02595275], dtype=float32),\n", " array([0.02479677], dtype=float32),\n", " array([0.0237], dtype=float32),\n", " array([0.02265339], dtype=float32),\n", " array([0.02167513], dtype=float32),\n", " array([0.02077813], dtype=float32),\n", " array([0.01996211], dtype=float32),\n", " array([0.0191979], dtype=float32),\n", " array([0.01846334], dtype=float32),\n", " array([0.01776404], dtype=float32),\n", " array([0.01710169], dtype=float32),\n", " array([0.01647772], dtype=float32),\n", " array([0.01589418], dtype=float32),\n", " array([0.0153296], dtype=float32),\n", " array([0.01480919], dtype=float32),\n", " array([0.01431813], dtype=float32),\n", " array([0.01384668], dtype=float32),\n", " array([0.01341317], dtype=float32),\n", " array([0.01301116], dtype=float32),\n", " array([0.01262761], dtype=float32),\n", " array([0.01226609], dtype=float32),\n", " array([0.01193028], dtype=float32),\n", " array([0.01162518], dtype=float32),\n", " array([0.01132936], dtype=float32),\n", " array([0.01103591], dtype=float32),\n", " array([0.0107461], dtype=float32),\n", " array([0.01046892], dtype=float32),\n", " array([0.01020851], dtype=float32),\n", " array([0.00995685], dtype=float32),\n", " array([0.00970961], dtype=float32),\n", " array([0.00949461], dtype=float32),\n", " array([0.00930674], dtype=float32),\n", " array([0.0091188], dtype=float32),\n", " array([0.00894598], dtype=float32),\n", " array([0.00876783], dtype=float32),\n", " array([0.00858588], dtype=float32),\n", " array([0.00839955], dtype=float32),\n", " array([0.00821982], dtype=float32),\n", " array([0.00804709], dtype=float32),\n", " array([0.00787459], dtype=float32),\n", " array([0.00771036], dtype=float32),\n", " array([0.00754671], dtype=float32),\n", " array([0.00740416], dtype=float32),\n", " array([0.00725957], dtype=float32),\n", " array([0.00711691], dtype=float32),\n", " array([0.00697032], dtype=float32),\n", " array([0.00683032], dtype=float32),\n", " array([0.0066928], dtype=float32),\n", " array([0.00656939], dtype=float32),\n", " array([0.00646281], dtype=float32),\n", " array([0.00635776], dtype=float32),\n", " array([0.00625345], dtype=float32),\n", " array([0.00614849], dtype=float32),\n", " array([0.00604802], dtype=float32),\n", " array([0.00594663], dtype=float32),\n", " array([0.00584267], dtype=float32),\n", " array([0.00573586], dtype=float32),\n", " array([0.00564113], dtype=float32),\n", " array([0.00555611], dtype=float32),\n", " array([0.00546922], dtype=float32),\n", " array([0.00538022], dtype=float32),\n", " array([0.00529412], dtype=float32),\n", " array([0.0052038], dtype=float32),\n", " array([0.00511062], dtype=float32),\n", " array([0.00501934], dtype=float32),\n", " array([0.00492939], dtype=float32),\n", " array([0.00484378], dtype=float32),\n", " array([0.0047608], dtype=float32),\n", " array([0.00467696], dtype=float32),\n", " array([0.00459164], dtype=float32),\n", " array([0.00450822], dtype=float32),\n", " array([0.00442378], dtype=float32),\n", " array([0.00434016], dtype=float32),\n", " array([0.00425913], dtype=float32),\n", " array([0.00418144], dtype=float32),\n", " array([0.00410846], dtype=float32),\n", " array([0.00403432], dtype=float32),\n", " array([0.00396164], dtype=float32),\n", " array([0.00388966], dtype=float32),\n", " array([0.00381772], dtype=float32),\n", " array([0.00374668], dtype=float32),\n", " array([0.00367686], dtype=float32),\n", " array([0.00361116], dtype=float32),\n", " array([0.00355041], dtype=float32),\n", " array([0.00349828], dtype=float32),\n", " array([0.00344607], dtype=float32),\n", " array([0.00339907], dtype=float32),\n", " array([0.0033515], dtype=float32),\n", " array([0.0033067], dtype=float32),\n", " array([0.0032599], dtype=float32),\n", " array([0.0032117], dtype=float32),\n", " array([0.00317386], dtype=float32),\n", " array([0.00313674], dtype=float32),\n", " array([0.00309983], dtype=float32),\n", " array([0.00306505], dtype=float32),\n", " array([0.00302898], dtype=float32),\n", " array([0.00299239], dtype=float32),\n", " array([0.00295825], dtype=float32),\n", " array([0.00292576], dtype=float32),\n", " array([0.00289435], dtype=float32),\n", " array([0.00286749], dtype=float32),\n", " array([0.002839], dtype=float32),\n", " array([0.00280783], dtype=float32),\n", " array([0.00278303], dtype=float32),\n", " array([0.00275987], dtype=float32),\n", " array([0.00273622], dtype=float32),\n", " array([0.0027116], dtype=float32),\n", " array([0.00268711], dtype=float32),\n", " array([0.00266542], dtype=float32),\n", " array([0.00264566], dtype=float32),\n", " array([0.00262384], dtype=float32),\n", " array([0.00260093], dtype=float32),\n", " array([0.0025769], dtype=float32),\n", " array([0.00255362], dtype=float32),\n", " array([0.00252923], dtype=float32),\n", " array([0.00250831], dtype=float32),\n", " array([0.00248595], dtype=float32),\n", " array([0.00246177], dtype=float32),\n", " array([0.00243588], dtype=float32),\n", " array([0.00241615], dtype=float32),\n", " array([0.00239742], dtype=float32),\n", " array([0.00238027], dtype=float32),\n", " array([0.00236432], dtype=float32),\n", " array([0.00234785], dtype=float32),\n", " array([0.00233038], dtype=float32),\n", " array([0.0023133], dtype=float32),\n", " array([0.00229659], dtype=float32),\n", " array([0.00228019], dtype=float32),\n", " array([0.00226641], dtype=float32),\n", " array([0.00225646], dtype=float32),\n", " array([0.00224488], dtype=float32),\n", " array([0.00223969], dtype=float32),\n", " array([0.00223436], dtype=float32),\n", " array([0.00222715], dtype=float32),\n", " array([0.00221926], dtype=float32),\n", " array([0.00221096], dtype=float32),\n", " array([0.00220209], dtype=float32),\n", " array([0.00219372], dtype=float32),\n", " array([0.00218328], dtype=float32),\n", " array([0.00217303], dtype=float32),\n", " array([0.00216029], dtype=float32),\n", " array([0.00214934], dtype=float32),\n", " array([0.00213939], dtype=float32),\n", " array([0.00213206], dtype=float32),\n", " array([0.00212225], dtype=float32),\n", " array([0.00211091], dtype=float32),\n", " array([0.00209845], dtype=float32),\n", " array([0.0020862], dtype=float32),\n", " array([0.00207273], dtype=float32),\n", " array([0.00206018], dtype=float32),\n", " array([0.00204672], dtype=float32),\n", " array([0.0020332], dtype=float32),\n", " array([0.00201869], dtype=float32),\n", " array([0.00200966], dtype=float32),\n", " array([0.0020023], dtype=float32),\n", " array([0.00199713], dtype=float32),\n", " array([0.00199067], dtype=float32),\n", " array([0.0019829], dtype=float32),\n", " array([0.00197719], dtype=float32),\n", " array([0.00196905], dtype=float32),\n", " array([0.00195856], dtype=float32),\n", " array([0.00194568], dtype=float32),\n", " array([0.00193084], dtype=float32),\n", " array([0.00191519], dtype=float32),\n", " array([0.00190115], dtype=float32),\n", " array([0.00188929], dtype=float32),\n", " array([0.00187574], dtype=float32),\n", " array([0.00186303], dtype=float32),\n", " array([0.00185313], dtype=float32),\n", " array([0.00184834], dtype=float32),\n", " array([0.00184312], dtype=float32),\n", " array([0.00184014], dtype=float32),\n", " array([0.00183669], dtype=float32),\n", " array([0.00183361], dtype=float32),\n", " array([0.00183377], dtype=float32),\n", " array([0.00183379], dtype=float32),\n", " array([0.00183388], dtype=float32),\n", " array([0.00183175], dtype=float32),\n", " array([0.00182798], dtype=float32),\n", " array([0.00182418], dtype=float32),\n", " array([0.00182125], dtype=float32),\n", " array([0.00181876], dtype=float32),\n", " array([0.00181563], dtype=float32),\n", " array([0.00181683], dtype=float32),\n", " array([0.00182146], dtype=float32),\n", " array([0.00182359], dtype=float32),\n", " array([0.00182237], dtype=float32),\n", " array([0.00181997], dtype=float32),\n", " array([0.00181531], dtype=float32),\n", " array([0.00180874], dtype=float32),\n", " array([0.00180134], dtype=float32),\n", " array([0.00179988], dtype=float32),\n", " array([0.00180631], dtype=float32),\n", " array([0.00181428], dtype=float32),\n", " array([0.00182256], dtype=float32),\n", " array([0.00182832], dtype=float32),\n", " array([0.00183132], dtype=float32),\n", " array([0.00183054], dtype=float32),\n", " array([0.00183423], dtype=float32),\n", " array([0.00183786], dtype=float32),\n", " array([0.00183919], dtype=float32),\n", " array([0.00183744], dtype=float32),\n", " array([0.00183271], dtype=float32),\n", " array([0.00182751], dtype=float32),\n", " array([0.00182456], dtype=float32),\n", " array([0.00182244], dtype=float32),\n", " array([0.0018205], dtype=float32),\n", " array([0.00181652], dtype=float32),\n", " array([0.00181679], dtype=float32),\n", " array([0.00182393], dtype=float32),\n", " array([0.00182918], dtype=float32),\n", " array([0.00183294], dtype=float32),\n", " array([0.00183439], dtype=float32),\n", " array([0.00183336], dtype=float32),\n", " array([0.00183369], dtype=float32),\n", " array([0.00183171], dtype=float32),\n", " array([0.00182747], dtype=float32),\n", " array([0.00182704], dtype=float32),\n", " array([0.00183075], dtype=float32),\n", " array([0.00183508], dtype=float32),\n", " array([0.00184318], dtype=float32),\n", " array([0.00185059], dtype=float32),\n", " array([0.00185523], dtype=float32),\n", " array([0.00185663], dtype=float32),\n", " array([0.00185422], dtype=float32),\n", " array([0.00185092], dtype=float32),\n", " array([0.00184496], dtype=float32),\n", " array([0.00183972], dtype=float32),\n", " array([0.00183422], dtype=float32),\n", " array([0.00182542], dtype=float32),\n", " array([0.00181576], dtype=float32),\n", " array([0.00180334], dtype=float32),\n", " array([0.00179152], dtype=float32),\n", " array([0.00178413], dtype=float32),\n", " array([0.00177527], dtype=float32),\n", " array([0.00177275], dtype=float32),\n", " array([0.0017733], dtype=float32),\n", " array([0.00177358], dtype=float32),\n", " array([0.00177023], dtype=float32),\n", " array([0.00176281], dtype=float32),\n", " array([0.00175413], dtype=float32),\n", " array([0.00174655], dtype=float32),\n", " array([0.00174192], dtype=float32),\n", " array([0.00173934], dtype=float32),\n", " array([0.00173713], dtype=float32),\n", " array([0.00173962], dtype=float32),\n", " array([0.00174911], dtype=float32),\n", " array([0.00175627], dtype=float32),\n", " array([0.00176019], dtype=float32),\n", " array([0.00176313], dtype=float32),\n", " array([0.00176398], dtype=float32),\n", " array([0.0017683], dtype=float32),\n", " array([0.00177455], dtype=float32),\n", " array([0.00177735], dtype=float32),\n", " array([0.00178031], dtype=float32),\n", " array([0.00178305], dtype=float32),\n", " array([0.00178802], dtype=float32),\n", " array([0.00179372], dtype=float32),\n", " array([0.00179653], dtype=float32),\n", " array([0.00179575], dtype=float32),\n", " array([0.0017955], dtype=float32),\n", " array([0.00179813], dtype=float32),\n", " array([0.00180457], dtype=float32),\n", " array([0.00181303], dtype=float32),\n", " array([0.00181834], dtype=float32),\n", " array([0.00182324], dtype=float32),\n", " array([0.00182519], dtype=float32),\n", " array([0.00182453], dtype=float32),\n", " array([0.00182047], dtype=float32),\n", " array([0.00181479], dtype=float32),\n", " array([0.00180858], dtype=float32),\n", " array([0.00180114], dtype=float32),\n", " array([0.00179145], dtype=float32),\n", " array([0.00178069], dtype=float32),\n", " array([0.00177254], dtype=float32),\n", " array([0.00176276], dtype=float32),\n", " array([0.00175269], dtype=float32),\n", " array([0.00174288], dtype=float32),\n", " array([0.00173395], dtype=float32),\n", " array([0.00172699], dtype=float32),\n", " array([0.00171898], dtype=float32),\n", " array([0.00170984], dtype=float32),\n", " array([0.00170344], dtype=float32),\n", " array([0.00169743], dtype=float32),\n", " array([0.00168992], dtype=float32),\n", " array([0.00168], dtype=float32),\n", " array([0.00166857], dtype=float32),\n", " array([0.00165685], dtype=float32),\n", " array([0.00164687], dtype=float32),\n", " array([0.00164652], dtype=float32),\n", " array([0.00164623], dtype=float32),\n", " array([0.00164269], dtype=float32),\n", " array([0.00163635], dtype=float32),\n", " array([0.00163017], dtype=float32),\n", " array([0.00162228], dtype=float32),\n", " array([0.00161524], dtype=float32),\n", " array([0.00160722], dtype=float32),\n", " array([0.00160193], dtype=float32),\n", " array([0.00159853], dtype=float32),\n", " array([0.00159817], dtype=float32),\n", " array([0.0016], dtype=float32),\n", " array([0.00160067], dtype=float32),\n", " array([0.0016052], dtype=float32),\n", " array([0.00160956], dtype=float32),\n", " array([0.00161166], dtype=float32),\n", " array([0.00161021], dtype=float32),\n", " array([0.00160516], dtype=float32),\n", " array([0.00159804], dtype=float32),\n", " array([0.00159209], dtype=float32),\n", " array([0.001586], dtype=float32),\n", " array([0.0015805], dtype=float32),\n", " array([0.00157572], dtype=float32),\n", " array([0.00157014], dtype=float32),\n", " array([0.00156988], dtype=float32),\n", " array([0.00157329], dtype=float32),\n", " array([0.0015738], dtype=float32),\n", " array([0.00157279], dtype=float32),\n", " array([0.00157064], dtype=float32),\n", " array([0.00156781], dtype=float32),\n", " array([0.00156759], dtype=float32),\n", " array([0.00156848], dtype=float32),\n", " array([0.00156828], dtype=float32),\n", " array([0.00156679], dtype=float32),\n", " array([0.00156294], dtype=float32),\n", " array([0.00155595], dtype=float32),\n", " array([0.00154738], dtype=float32),\n", " array([0.00153798], dtype=float32),\n", " array([0.001526], dtype=float32),\n", " array([0.00151203], dtype=float32),\n", " array([0.00149792], dtype=float32),\n", " array([0.00148507], dtype=float32),\n", " array([0.00147201], dtype=float32),\n", " array([0.00145875], dtype=float32),\n", " array([0.00144604], dtype=float32),\n", " array([0.00143205], dtype=float32),\n", " array([0.00141676], dtype=float32),\n", " array([0.00140255], dtype=float32),\n", " array([0.00139512], dtype=float32),\n", " array([0.00139199], dtype=float32),\n", " array([0.00139086], dtype=float32),\n", " array([0.00138944], dtype=float32),\n", " array([0.00138875], dtype=float32),\n", " array([0.00139053], dtype=float32),\n", " array([0.00139379], dtype=float32),\n", " array([0.00140053], dtype=float32),\n", " array([0.00140847], dtype=float32),\n", " array([0.00141433], dtype=float32),\n", " array([0.00141747], dtype=float32),\n", " array([0.00142161], dtype=float32),\n", " array([0.00142876], dtype=float32),\n", " array([0.00143639], dtype=float32),\n", " array([0.00144372], dtype=float32),\n", " array([0.0014485], dtype=float32),\n", " array([0.00145095], dtype=float32),\n", " array([0.00145206], dtype=float32),\n", " array([0.0014555], dtype=float32),\n", " array([0.00145948], dtype=float32),\n", " array([0.00146184], dtype=float32),\n", " array([0.00146241], dtype=float32),\n", " array([0.001464], dtype=float32),\n", " array([0.00146559], dtype=float32),\n", " array([0.00146732], dtype=float32),\n", " array([0.00146733], dtype=float32),\n", " array([0.00146663], dtype=float32),\n", " array([0.00146729], dtype=float32),\n", " array([0.00146836], dtype=float32),\n", " array([0.00146883], dtype=float32),\n", " array([0.00147034], dtype=float32),\n", " array([0.00147439], dtype=float32),\n", " array([0.00148029], dtype=float32),\n", " array([0.00148456], dtype=float32),\n", " array([0.0014865], dtype=float32),\n", " array([0.00148679], dtype=float32),\n", " array([0.00148676], dtype=float32),\n", " array([0.00148551], dtype=float32),\n", " array([0.00148319], dtype=float32),\n", " array([0.00148178], dtype=float32),\n", " array([0.00147944], dtype=float32),\n", " array([0.00147406], dtype=float32),\n", " array([0.00146826], dtype=float32),\n", " array([0.0014638], dtype=float32),\n", " array([0.00146282], dtype=float32),\n", " array([0.00146818], dtype=float32),\n", " array([0.00147576], dtype=float32),\n", " array([0.00148512], dtype=float32),\n", " array([0.00150254], dtype=float32),\n", " array([0.0015235], dtype=float32),\n", " array([0.00154413], dtype=float32),\n", " array([0.0015639], dtype=float32),\n", " array([0.00157994], dtype=float32),\n", " array([0.00159106], dtype=float32),\n", " array([0.00159657], dtype=float32),\n", " array([0.00159705], dtype=float32),\n", " array([0.00159394], dtype=float32),\n", " array([0.00158951], dtype=float32),\n", " array([0.00158767], dtype=float32),\n", " array([0.00159235], dtype=float32),\n", " array([0.00160518], dtype=float32),\n", " array([0.00162063], dtype=float32),\n", " array([0.00163526], dtype=float32),\n", " array([0.00165115], dtype=float32),\n", " array([0.00166538], dtype=float32),\n", " array([0.00167877], dtype=float32),\n", " array([0.00169198], dtype=float32),\n", " array([0.00170053], dtype=float32),\n", " array([0.00170503], dtype=float32),\n", " array([0.00170874], dtype=float32),\n", " array([0.00170961], dtype=float32),\n", " array([0.00170763], dtype=float32),\n", " array([0.00170562], dtype=float32),\n", " array([0.00170381], dtype=float32),\n", " array([0.00170173], dtype=float32),\n", " array([0.0016989], dtype=float32),\n", " array([0.00169579], dtype=float32),\n", " array([0.00169018], dtype=float32),\n", " array([0.00168042], dtype=float32),\n", " array([0.00166797], dtype=float32),\n", " array([0.001654], dtype=float32),\n", " array([0.00163894], dtype=float32),\n", " array([0.00162541], dtype=float32),\n", " array([0.00161341], dtype=float32),\n", " array([0.00160067], dtype=float32),\n", " array([0.0015862], dtype=float32),\n", " array([0.00156989], dtype=float32),\n", " array([0.00155158], dtype=float32),\n", " array([0.00153244], dtype=float32),\n", " array([0.00151446], dtype=float32),\n", " array([0.00149939], dtype=float32),\n", " array([0.00148594], dtype=float32),\n", " array([0.00147125], dtype=float32),\n", " array([0.00145503], dtype=float32),\n", " array([0.00143839], dtype=float32),\n", " array([0.00142231], dtype=float32),\n", " array([0.00140844], dtype=float32),\n", " array([0.00139684], dtype=float32),\n", " array([0.00138598], dtype=float32),\n", " array([0.00137633], dtype=float32),\n", " array([0.00136671], dtype=float32),\n", " array([0.00135522], dtype=float32),\n", " array([0.00134381], dtype=float32),\n", " array([0.00133418], dtype=float32),\n", " array([0.00132563], dtype=float32),\n", " array([0.00131941], dtype=float32),\n", " array([0.00131509], dtype=float32),\n", " array([0.00131141], dtype=float32),\n", " array([0.00130855], dtype=float32),\n", " array([0.00130588], dtype=float32),\n", " array([0.00130106], dtype=float32),\n", " array([0.00129328], dtype=float32),\n", " array([0.00128461], dtype=float32),\n", " array([0.0012769], dtype=float32),\n", " array([0.00127011], dtype=float32),\n", " array([0.00126437], dtype=float32),\n", " array([0.00125901], dtype=float32),\n", " array([0.00125334], dtype=float32),\n", " array([0.00124673], dtype=float32),\n", " array([0.00123926], dtype=float32),\n", " array([0.00123087], dtype=float32),\n", " array([0.00122134], dtype=float32),\n", " array([0.00121082], dtype=float32),\n", " array([0.0011995], dtype=float32),\n", " array([0.00118711], dtype=float32),\n", " array([0.00117349], dtype=float32),\n", " array([0.00115941], dtype=float32),\n", " array([0.00114587], dtype=float32),\n", " array([0.00113269], dtype=float32),\n", " array([0.00112079], dtype=float32),\n", " array([0.00111576], dtype=float32),\n", " array([0.00111758], dtype=float32),\n", " array([0.00111998], dtype=float32),\n", " array([0.00112113], dtype=float32),\n", " array([0.00112198], dtype=float32),\n", " array([0.00112397], dtype=float32),\n", " array([0.00112642], dtype=float32),\n", " array([0.00112843], dtype=float32),\n", " array([0.0011298], dtype=float32),\n", " array([0.00113019], dtype=float32),\n", " array([0.0011312], dtype=float32),\n", " array([0.00113428], dtype=float32),\n", " array([0.00113903], dtype=float32),\n", " array([0.00114429], dtype=float32),\n", " array([0.00114918], dtype=float32),\n", " array([0.00115256], dtype=float32),\n", " array([0.00115395], dtype=float32),\n", " array([0.00115332], dtype=float32),\n", " array([0.00115059], dtype=float32),\n", " array([0.00114702], dtype=float32),\n", " array([0.00114464], dtype=float32),\n", " array([0.00114366], dtype=float32),\n", " array([0.00114327], dtype=float32),\n", " array([0.00114298], dtype=float32),\n", " array([0.00114151], dtype=float32),\n", " array([0.00113832], dtype=float32),\n", " array([0.0011344], dtype=float32),\n", " array([0.0011301], dtype=float32),\n", " array([0.00112474], dtype=float32),\n", " array([0.00111822], dtype=float32),\n", " array([0.00111088], dtype=float32),\n", " array([0.00110275], dtype=float32),\n", " array([0.00109373], dtype=float32),\n", " array([0.00108424], dtype=float32),\n", " array([0.00107515], dtype=float32),\n", " array([0.00106749], dtype=float32),\n", " array([0.0010618], dtype=float32),\n", " array([0.00105741], dtype=float32),\n", " array([0.0010533], dtype=float32),\n", " array([0.00104905], dtype=float32),\n", " array([0.00104462], dtype=float32),\n", " array([0.00104033], dtype=float32),\n", " array([0.00103601], dtype=float32),\n", " array([0.00103129], dtype=float32),\n", " array([0.00102727], dtype=float32),\n", " array([0.00102487], dtype=float32),\n", " array([0.00102315], dtype=float32),\n", " array([0.00102107], dtype=float32),\n", " array([0.00101894], dtype=float32),\n", " array([0.0010182], dtype=float32),\n", " array([0.00101997], dtype=float32),\n", " array([0.00102315], dtype=float32),\n", " array([0.00102598], dtype=float32),\n", " array([0.00102794], dtype=float32),\n", " array([0.00102943], dtype=float32),\n", " array([0.00103236], dtype=float32),\n", " array([0.00103829], dtype=float32),\n", " array([0.00104664], dtype=float32),\n", " array([0.00105607], dtype=float32),\n", " array([0.00106594], dtype=float32),\n", " array([0.00107563], dtype=float32),\n", " array([0.0010843], dtype=float32),\n", " array([0.00109252], dtype=float32),\n", " array([0.0011015], dtype=float32),\n", " array([0.00111197], dtype=float32),\n", " array([0.00112399], dtype=float32),\n", " array([0.00113561], dtype=float32),\n", " array([0.0011445], dtype=float32),\n", " array([0.00115022], dtype=float32),\n", " array([0.00115445], dtype=float32),\n", " array([0.00116161], dtype=float32),\n", " array([0.00117374], dtype=float32),\n", " array([0.00118705], dtype=float32),\n", " array([0.00119818], dtype=float32),\n", " array([0.00120766], dtype=float32),\n", " array([0.00121609], dtype=float32),\n", " array([0.00122249], dtype=float32),\n", " array([0.00122641], dtype=float32),\n", " array([0.00122879], dtype=float32),\n", " array([0.00123056], dtype=float32),\n", " array([0.00123214], dtype=float32),\n", " array([0.00123364], dtype=float32),\n", " array([0.00123556], dtype=float32),\n", " array([0.00124014], dtype=float32),\n", " array([0.00125085], dtype=float32),\n", " array([0.00126676], dtype=float32),\n", " array([0.00128248], dtype=float32),\n", " array([0.00129553], dtype=float32),\n", " array([0.00130671], dtype=float32),\n", " array([0.00131626], dtype=float32),\n", " array([0.00132325], dtype=float32),\n", " array([0.00132672], dtype=float32),\n", " array([0.00132712], dtype=float32),\n", " array([0.00132697], dtype=float32),\n", " array([0.00132896], dtype=float32),\n", " array([0.00133434], dtype=float32),\n", " array([0.00134297], dtype=float32),\n", " array([0.00135363], dtype=float32),\n", " array([0.0013652], dtype=float32),\n", " array([0.00137838], dtype=float32),\n", " array([0.00139325], dtype=float32),\n", " array([0.00140675], dtype=float32),\n", " array([0.00141621], dtype=float32),\n", " array([0.00142132], dtype=float32),\n", " array([0.00142228], dtype=float32),\n", " array([0.00141996], dtype=float32),\n", " array([0.00141651], dtype=float32),\n", " array([0.00141536], dtype=float32),\n", " array([0.00141773], dtype=float32),\n", " array([0.00142176], dtype=float32),\n", " array([0.00142502], dtype=float32),\n", " array([0.0014267], dtype=float32),\n", " array([0.0014269], dtype=float32),\n", " array([0.00142657], dtype=float32),\n", " array([0.00142574], dtype=float32),\n", " array([0.00142297], dtype=float32),\n", " array([0.00141743], dtype=float32),\n", " array([0.00141016], dtype=float32),\n", " array([0.00140439], dtype=float32),\n", " array([0.00140209], dtype=float32),\n", " array([0.00140139], dtype=float32),\n", " array([0.00140032], dtype=float32),\n", " array([0.00140045], dtype=float32),\n", " array([0.0014025], dtype=float32),\n", " array([0.00140422], dtype=float32),\n", " array([0.00140301], dtype=float32),\n", " array([0.00139784], dtype=float32),\n", " array([0.00138924], dtype=float32),\n", " array([0.00137902], dtype=float32),\n", " array([0.00136934], dtype=float32),\n", " array([0.00136089], dtype=float32),\n", " array([0.00135243], dtype=float32),\n", " array([0.00134332], dtype=float32),\n", " array([0.00133559], dtype=float32),\n", " array([0.00133024], dtype=float32),\n", " array([0.00132513], dtype=float32),\n", " array([0.00131791], dtype=float32),\n", " array([0.0013089], dtype=float32),\n", " array([0.00129894], dtype=float32),\n", " array([0.00128803], dtype=float32),\n", " array([0.00127588], dtype=float32),\n", " array([0.00126243], dtype=float32),\n", " array([0.0012475], dtype=float32),\n", " array([0.00123127], dtype=float32),\n", " array([0.00121472], dtype=float32),\n", " array([0.00119896], dtype=float32),\n", " array([0.00118423], dtype=float32),\n", " array([0.00116994], dtype=float32),\n", " array([0.00115554], dtype=float32),\n", " array([0.00114165], dtype=float32),\n", " array([0.00112959], dtype=float32),\n", " array([0.00112002], dtype=float32),\n", " array([0.00111279], dtype=float32),\n", " array([0.00110712], dtype=float32),\n", " array([0.00110134], dtype=float32),\n", " array([0.00109385], dtype=float32),\n", " array([0.00108436], dtype=float32),\n", " array([0.00107371], dtype=float32),\n", " array([0.00106282], dtype=float32),\n", " array([0.00105171], dtype=float32),\n", " array([0.00103994], dtype=float32),\n", " array([0.00102787], dtype=float32),\n", " array([0.0010173], dtype=float32),\n", " array([0.00101013], dtype=float32),\n", " array([0.00100634], dtype=float32),\n", " array([0.00100433], dtype=float32),\n", " array([0.00100274], dtype=float32),\n", " array([0.00100113], dtype=float32),\n", " array([0.00099923], dtype=float32),\n", " array([0.00099649], dtype=float32),\n", " array([0.00099237], dtype=float32),\n", " array([0.00098705], dtype=float32),\n", " array([0.00098139], dtype=float32),\n", " array([0.00097669], dtype=float32),\n", " array([0.00097363], dtype=float32),\n", " array([0.00097169], dtype=float32),\n", " array([0.00096981], dtype=float32),\n", " array([0.00096743], dtype=float32),\n", " array([0.0009653], dtype=float32),\n", " array([0.0009647], dtype=float32),\n", " array([0.0009654], dtype=float32),\n", " array([0.00096575], dtype=float32),\n", " array([0.00096496], dtype=float32),\n", " array([0.00096367], dtype=float32),\n", " array([0.00096243], dtype=float32),\n", " array([0.00096136], dtype=float32),\n", " array([0.0009611], dtype=float32),\n", " array([0.00096294], dtype=float32),\n", " array([0.0009677], dtype=float32),\n", " array([0.00097429], dtype=float32),\n", " array([0.00098047], dtype=float32),\n", " array([0.00098474], dtype=float32),\n", " array([0.0009868], dtype=float32),\n", " array([0.00098703], dtype=float32),\n", " array([0.00098633], dtype=float32),\n", " array([0.00098526], dtype=float32),\n", " array([0.0009839], dtype=float32),\n", " array([0.00098231], dtype=float32),\n", " array([0.00098039], dtype=float32),\n", " array([0.000978], dtype=float32),\n", " array([0.00097534], dtype=float32),\n", " array([0.00097327], dtype=float32),\n", " array([0.00097318], dtype=float32),\n", " array([0.00097572], dtype=float32),\n", " array([0.00098054], dtype=float32),\n", " array([0.00098704], dtype=float32),\n", " array([0.0009944], dtype=float32),\n", " array([0.001002], dtype=float32),\n", " array([0.0010099], dtype=float32),\n", " array([0.00101808], dtype=float32),\n", " array([0.00102594], dtype=float32),\n", " array([0.00103326], dtype=float32),\n", " array([0.00104067], dtype=float32),\n", " array([0.00104853], dtype=float32),\n", " array([0.00105593], dtype=float32),\n", " array([0.00106152], dtype=float32),\n", " array([0.00106506], dtype=float32),\n", " array([0.00106754], dtype=float32),\n", " array([0.00107014], dtype=float32),\n", " array([0.00107301], dtype=float32),\n", " array([0.00107529], dtype=float32),\n", " array([0.00107644], dtype=float32),\n", " array([0.00107685], dtype=float32),\n", " array([0.00107683], dtype=float32),\n", " array([0.00107591], dtype=float32),\n", " array([0.00107336], dtype=float32),\n", " array([0.00106896], dtype=float32),\n", " array([0.00106392], dtype=float32),\n", " array([0.00106135], dtype=float32),\n", " array([0.00106367], dtype=float32),\n", " array([0.00107001], dtype=float32),\n", " array([0.00107701], dtype=float32),\n", " array([0.0010823], dtype=float32),\n", " array([0.0010858], dtype=float32),\n", " array([0.00108816], dtype=float32),\n", " array([0.00108956], dtype=float32),\n", " array([0.00109003], dtype=float32),\n", " array([0.00109058], dtype=float32),\n", " array([0.0010929], dtype=float32),\n", " array([0.00109769], dtype=float32),\n", " array([0.00110413], dtype=float32),\n", " array([0.00111057], dtype=float32),\n", " array([0.00111556], dtype=float32),\n", " array([0.00111824], dtype=float32),\n", " array([0.00111856], dtype=float32),\n", " array([0.00111725], dtype=float32),\n", " array([0.00111541], dtype=float32),\n", " array([0.00111369], dtype=float32),\n", " array([0.00111196], dtype=float32),\n", " array([0.00110957], dtype=float32),\n", " array([0.0011061], dtype=float32),\n", " array([0.00110225], dtype=float32),\n", " array([0.00109932], dtype=float32),\n", " array([0.00109772], dtype=float32),\n", " array([0.00109633], dtype=float32),\n", " array([0.00109368], dtype=float32),\n", " array([0.00108911], dtype=float32),\n", " array([0.00108287], dtype=float32),\n", " array([0.00107569], dtype=float32),\n", " array([0.0010682], dtype=float32),\n", " array([0.00106042], dtype=float32),\n", " array([0.00105211], dtype=float32),\n", " array([0.00104344], dtype=float32),\n", " array([0.00103467], dtype=float32),\n", " array([0.0010255], dtype=float32),\n", " array([0.00101528], dtype=float32),\n", " array([0.00100366], dtype=float32),\n", " array([0.00099087], dtype=float32),\n", " array([0.00097754], dtype=float32),\n", " array([0.00096431], dtype=float32),\n", " array([0.00095152], dtype=float32),\n", " array([0.00093899], dtype=float32),\n", " array([0.00092646], dtype=float32),\n", " array([0.00091383], dtype=float32),\n", " array([0.00090113], dtype=float32),\n", " array([0.00088843], dtype=float32),\n", " array([0.00087625], dtype=float32),\n", " array([0.00086597], dtype=float32),\n", " array([0.00085924], dtype=float32),\n", " array([0.00085692], dtype=float32),\n", " array([0.00085874], dtype=float32),\n", " array([0.00086344], dtype=float32),\n", " array([0.00086916], dtype=float32),\n", " array([0.00087429], dtype=float32),\n", " array([0.00087835], dtype=float32),\n", " array([0.00088158], dtype=float32),\n", " array([0.00088397], dtype=float32),\n", " array([0.00088515], dtype=float32),\n", " array([0.00088481], dtype=float32),\n", " array([0.00088294], dtype=float32),\n", " array([0.00088021], dtype=float32),\n", " array([0.00087813], dtype=float32),\n", " array([0.00087796], dtype=float32),\n", " array([0.00087982], dtype=float32),\n", " array([0.00088326], dtype=float32),\n", " array([0.00088827], dtype=float32),\n", " array([0.00089539], dtype=float32),\n", " array([0.00090517], dtype=float32),\n", " array([0.00091757], dtype=float32),\n", " array([0.00093172], dtype=float32),\n", " array([0.00094615], dtype=float32),\n", " array([0.00095995], dtype=float32),\n", " array([0.00097357], dtype=float32),\n", " array([0.00098804], dtype=float32),\n", " array([0.00100398], dtype=float32),\n", " array([0.00102127], dtype=float32),\n", " array([0.00103909], dtype=float32),\n", " array([0.00105626], dtype=float32),\n", " array([0.00107161], dtype=float32),\n", " array([0.00108454], dtype=float32),\n", " array([0.00109509], dtype=float32),\n", " array([0.00110404], dtype=float32),\n", " array([0.00111305], dtype=float32),\n", " array([0.00112386], dtype=float32),\n", " array([0.00113707], dtype=float32),\n", " array([0.00115188], dtype=float32),\n", " array([0.00116771], dtype=float32),\n", " array([0.00118543], dtype=float32),\n", " array([0.00120608], dtype=float32),\n", " array([0.00122909], dtype=float32),\n", " array([0.00125255], dtype=float32),\n", " array([0.00127513], dtype=float32),\n", " array([0.00129668], dtype=float32),\n", " array([0.00131727], dtype=float32),\n", " array([0.00133615], dtype=float32),\n", " array([0.00135182], dtype=float32),\n", " array([0.00136287], dtype=float32),\n", " array([0.00136876], dtype=float32),\n", " array([0.00137045], dtype=float32),\n", " array([0.00136967], dtype=float32),\n", " array([0.00136763], dtype=float32),\n", " array([0.00136438], dtype=float32),\n", " array([0.00135951], dtype=float32),\n", " array([0.00135358], dtype=float32),\n", " array([0.00134908], dtype=float32),\n", " array([0.00134941], dtype=float32),\n", " array([0.00135634], dtype=float32),\n", " array([0.00136884], dtype=float32),\n", " array([0.00138412], dtype=float32),\n", " array([0.00140013], dtype=float32),\n", " array([0.00141705], dtype=float32),\n", " array([0.00143558], dtype=float32),\n", " array([0.0014548], dtype=float32),\n", " array([0.00147221], dtype=float32),\n", " array([0.00148541], dtype=float32),\n", " array([0.0014946], dtype=float32),\n", " array([0.00150287], dtype=float32),\n", " array([0.00151351], dtype=float32),\n", " array([0.00152752], dtype=float32),\n", " array([0.00154332], dtype=float32),\n", " array([0.00155805], dtype=float32),\n", " array([0.00156988], dtype=float32),\n", " array([0.00157898], dtype=float32),\n", " array([0.00158728], dtype=float32),\n", " array([0.00159694], dtype=float32),\n", " array([0.00160863], dtype=float32),\n", " array([0.00162136], dtype=float32),\n", " array([0.00163295], dtype=float32),\n", " array([0.00164108], dtype=float32),\n", " array([0.00164424], dtype=float32),\n", " array([0.0016416], dtype=float32),\n", " array([0.00163315], dtype=float32),\n", " array([0.00162061], dtype=float32),\n", " array([0.00160729], dtype=float32),\n", " array([0.00159641], dtype=float32),\n", " array([0.00158941], dtype=float32),\n", " array([0.00158569], dtype=float32),\n", " array([0.00158385], dtype=float32),\n", " array([0.0015831], dtype=float32),\n", " array([0.00158376], dtype=float32),\n", " array([0.00158672], dtype=float32),\n", " array([0.00159256], dtype=float32),\n", " array([0.00160139], dtype=float32),\n", " array([0.00161363], dtype=float32),\n", " array([0.00163085], dtype=float32),\n", " array([0.0016555], dtype=float32),\n", " array([0.00168956], dtype=float32),\n", " array([0.00173302], dtype=float32),\n", " array([0.00178323], dtype=float32),\n", " array([0.00183599], dtype=float32),\n", " array([0.00188691], dtype=float32),\n", " array([0.00193216], dtype=float32),\n", " array([0.0019689], dtype=float32),\n", " array([0.0019949], dtype=float32),\n", " array([0.00200837], dtype=float32),\n", " array([0.00200874], dtype=float32),\n", " array([0.00199684], dtype=float32),\n", " array([0.00197458], dtype=float32),\n", " array([0.00194475], dtype=float32),\n", " array([0.00191047], dtype=float32),\n", " array([0.00187438], dtype=float32),\n", " array([0.00183834], dtype=float32),\n", " array([0.00180379], dtype=float32),\n", " array([0.0017719], dtype=float32),\n", " array([0.0017436], dtype=float32),\n", " array([0.00171934], dtype=float32),\n", " array([0.00169895], dtype=float32),\n", " array([0.00168193], dtype=float32),\n", " array([0.00166761], dtype=float32),\n", " array([0.00165511], dtype=float32),\n", " array([0.0016435], dtype=float32),\n", " array([0.00163181], dtype=float32),\n", " array([0.00161918], dtype=float32),\n", " array([0.00160511], dtype=float32),\n", " array([0.0015893], dtype=float32),\n", " array([0.0015717], dtype=float32),\n", " array([0.00155247], dtype=float32),\n", " array([0.00153184], dtype=float32),\n", " array([0.00150996], dtype=float32),\n", " array([0.00148692], dtype=float32),\n", " array([0.00146293], dtype=float32),\n", " array([0.00143836], dtype=float32),\n", " array([0.00141373], dtype=float32),\n", " array([0.00138953], dtype=float32),\n", " array([0.00136591], dtype=float32),\n", " array([0.00134261], dtype=float32),\n", " array([0.0013194], dtype=float32),\n", " array([0.00129636], dtype=float32),\n", " array([0.00127389], dtype=float32),\n", " array([0.00125246], dtype=float32),\n", " array([0.00123238], dtype=float32),\n", " array([0.00121365], dtype=float32),\n", " array([0.00119593], dtype=float32),\n", " array([0.0011788], dtype=float32),\n", " array([0.00116193], dtype=float32),\n", " array([0.00114508], dtype=float32),\n", " array([0.00112821], dtype=float32),\n", " array([0.00111169], dtype=float32),\n", " array([0.00109641], dtype=float32),\n", " array([0.001084], dtype=float32),\n", " array([0.00107658], dtype=float32),\n", " array([0.00107596], dtype=float32),\n", " array([0.00108284], dtype=float32),\n", " array([0.00109642], dtype=float32),\n", " array([0.00111452], dtype=float32),\n", " array([0.00113423], dtype=float32),\n", " array([0.00115279], dtype=float32),\n", " array([0.0011682], dtype=float32),\n", " array([0.00117938], dtype=float32),\n", " array([0.00118625], dtype=float32),\n", " array([0.00118951], dtype=float32),\n", " array([0.00119028], dtype=float32),\n", " array([0.00118966], dtype=float32),\n", " array([0.0011885], dtype=float32),\n", " array([0.0011873], dtype=float32),\n", " array([0.00118628], dtype=float32),\n", " array([0.00118544], dtype=float32),\n", " array([0.00118453], dtype=float32),\n", " array([0.00118333], dtype=float32),\n", " array([0.00118191], dtype=float32),\n", " array([0.00118088], dtype=float32),\n", " array([0.0011812], dtype=float32),\n", " array([0.00118371], dtype=float32),\n", " array([0.00118856], dtype=float32),\n", " array([0.00119509], dtype=float32),\n", " array([0.00120211], dtype=float32),\n", " array([0.00120828], dtype=float32),\n", " array([0.00121246], dtype=float32),\n", " array([0.00121388], dtype=float32),\n", " array([0.00121226], dtype=float32),\n", " array([0.00120795], dtype=float32),\n", " array([0.0012019], dtype=float32),\n", " array([0.00119553], dtype=float32),\n", " array([0.00119023], dtype=float32),\n", " array([0.00118691], dtype=float32),\n", " array([0.00118559], dtype=float32),\n", " array([0.00118544], dtype=float32),\n", " array([0.00118521], dtype=float32),\n", " array([0.00118392], dtype=float32),\n", " array([0.00118129], dtype=float32),\n", " array([0.0011779], dtype=float32),\n", " array([0.00117493], dtype=float32),\n", " array([0.00117364], dtype=float32),\n", " array([0.00117488], dtype=float32),\n", " array([0.00117879], dtype=float32),\n", " array([0.00118493], dtype=float32),\n", " array([0.00119255], dtype=float32),\n", " array([0.00120107], dtype=float32),\n", " array([0.00121065], dtype=float32),\n", " array([0.00122238], dtype=float32),\n", " array([0.00123774], dtype=float32),\n", " array([0.00125791], dtype=float32),\n", " array([0.00128318], dtype=float32),\n", " array([0.00131264], dtype=float32),\n", " array([0.00134432], dtype=float32),\n", " ...],\n", " 'covar_module.mixture_means': [array([[[1.4098269]]], dtype=float32),\n", " array([[[1.4083471]]], dtype=float32),\n", " array([[[1.4069867]]], dtype=float32),\n", " array([[[1.405637]]], dtype=float32),\n", " array([[[1.4041992]]], dtype=float32),\n", " array([[[1.4027482]]], dtype=float32),\n", " array([[[1.4013164]]], dtype=float32),\n", " array([[[1.399896]]], dtype=float32),\n", " array([[[1.3984716]]], dtype=float32),\n", " array([[[1.3970507]]], dtype=float32),\n", " array([[[1.3956414]]], dtype=float32),\n", " array([[[1.3942391]]], dtype=float32),\n", " array([[[1.3928334]]], dtype=float32),\n", " array([[[1.391428]]], dtype=float32),\n", " array([[[1.3900336]]], dtype=float32),\n", " array([[[1.3886323]]], dtype=float32),\n", " array([[[1.387241]]], dtype=float32),\n", " array([[[1.385842]]], dtype=float32),\n", " array([[[1.3844556]]], dtype=float32),\n", " array([[[1.383082]]], dtype=float32),\n", " array([[[1.3817132]]], dtype=float32),\n", " array([[[1.3803208]]], dtype=float32),\n", " array([[[1.3789198]]], dtype=float32),\n", " array([[[1.3775338]]], dtype=float32),\n", " array([[[1.3761628]]], dtype=float32),\n", " array([[[1.3748033]]], dtype=float32),\n", " array([[[1.3734264]]], dtype=float32),\n", " array([[[1.3720505]]], dtype=float32),\n", " array([[[1.3706912]]], dtype=float32),\n", " array([[[1.3693291]]], dtype=float32),\n", " array([[[1.3679707]]], dtype=float32),\n", " array([[[1.3666152]]], dtype=float32),\n", " array([[[1.3652499]]], dtype=float32),\n", " array([[[1.3638897]]], dtype=float32),\n", " array([[[1.3625306]]], dtype=float32),\n", " array([[[1.3611768]]], dtype=float32),\n", " array([[[1.3598219]]], dtype=float32),\n", " array([[[1.358474]]], dtype=float32),\n", " array([[[1.3571205]]], dtype=float32),\n", " array([[[1.3557739]]], dtype=float32),\n", " array([[[1.3544197]]], dtype=float32),\n", " array([[[1.3530699]]], dtype=float32),\n", " array([[[1.3517256]]], dtype=float32),\n", " array([[[1.3503829]]], dtype=float32),\n", " array([[[1.3490505]]], dtype=float32),\n", " array([[[1.3477153]]], dtype=float32),\n", " array([[[1.3463802]]], dtype=float32),\n", " array([[[1.3450464]]], dtype=float32),\n", " array([[[1.3436992]]], dtype=float32),\n", " array([[[1.3423656]]], dtype=float32),\n", " array([[[1.3410329]]], dtype=float32),\n", " array([[[1.3396986]]], dtype=float32),\n", " array([[[1.3383634]]], dtype=float32),\n", " array([[[1.3370243]]], dtype=float32),\n", " array([[[1.3356779]]], dtype=float32),\n", " array([[[1.3343349]]], dtype=float32),\n", " array([[[1.3329977]]], dtype=float32),\n", " array([[[1.331661]]], dtype=float32),\n", " array([[[1.3303211]]], dtype=float32),\n", " array([[[1.3289895]]], dtype=float32),\n", " array([[[1.3276589]]], dtype=float32),\n", " array([[[1.3263314]]], dtype=float32),\n", " array([[[1.3249911]]], dtype=float32),\n", " array([[[1.3236669]]], dtype=float32),\n", " array([[[1.322344]]], dtype=float32),\n", " array([[[1.3210043]]], dtype=float32),\n", " array([[[1.3196911]]], dtype=float32),\n", " array([[[1.3183767]]], dtype=float32),\n", " array([[[1.3170707]]], dtype=float32),\n", " array([[[1.3157647]]], dtype=float32),\n", " array([[[1.314473]]], dtype=float32),\n", " array([[[1.3131758]]], dtype=float32),\n", " array([[[1.311878]]], dtype=float32),\n", " array([[[1.3105805]]], dtype=float32),\n", " array([[[1.3092694]]], dtype=float32),\n", " array([[[1.3079717]]], dtype=float32),\n", " array([[[1.3066695]]], dtype=float32),\n", " array([[[1.305363]]], dtype=float32),\n", " array([[[1.3040406]]], dtype=float32),\n", " array([[[1.3027229]]], dtype=float32),\n", " array([[[1.3014064]]], dtype=float32),\n", " array([[[1.3000804]]], dtype=float32),\n", " array([[[1.2987609]]], dtype=float32),\n", " array([[[1.2974483]]], dtype=float32),\n", " array([[[1.2961372]]], dtype=float32),\n", " array([[[1.2948287]]], dtype=float32),\n", " array([[[1.293519]]], dtype=float32),\n", " array([[[1.2922148]]], dtype=float32),\n", " array([[[1.290908]]], dtype=float32),\n", " array([[[1.2896101]]], dtype=float32),\n", " array([[[1.2882951]]], dtype=float32),\n", " array([[[1.2869885]]], dtype=float32),\n", " array([[[1.2856851]]], dtype=float32),\n", " array([[[1.2843839]]], dtype=float32),\n", " array([[[1.2830845]]], dtype=float32),\n", " array([[[1.2817824]]], dtype=float32),\n", " array([[[1.2804866]]], dtype=float32),\n", " array([[[1.2791976]]], dtype=float32),\n", " array([[[1.2779151]]], dtype=float32),\n", " array([[[1.2766341]]], dtype=float32),\n", " array([[[1.2753518]]], dtype=float32),\n", " array([[[1.2740805]]], dtype=float32),\n", " array([[[1.2728138]]], dtype=float32),\n", " array([[[1.2715487]]], dtype=float32),\n", " array([[[1.2702872]]], dtype=float32),\n", " array([[[1.2690386]]], dtype=float32),\n", " array([[[1.2677791]]], dtype=float32),\n", " array([[[1.2665184]]], dtype=float32),\n", " array([[[1.2652575]]], dtype=float32),\n", " array([[[1.2640004]]], dtype=float32),\n", " array([[[1.2627429]]], dtype=float32),\n", " array([[[1.2614855]]], dtype=float32),\n", " array([[[1.2602309]]], dtype=float32),\n", " array([[[1.2589742]]], dtype=float32),\n", " array([[[1.2577193]]], dtype=float32),\n", " array([[[1.2564634]]], dtype=float32),\n", " array([[[1.2552078]]], dtype=float32),\n", " array([[[1.2539531]]], dtype=float32),\n", " array([[[1.2527037]]], dtype=float32),\n", " array([[[1.251457]]], dtype=float32),\n", " array([[[1.2502081]]], dtype=float32),\n", " array([[[1.2489601]]], dtype=float32),\n", " array([[[1.2477081]]], dtype=float32),\n", " array([[[1.2464632]]], dtype=float32),\n", " array([[[1.2452195]]], dtype=float32),\n", " array([[[1.2439715]]], dtype=float32),\n", " array([[[1.2427251]]], dtype=float32),\n", " array([[[1.2414819]]], dtype=float32),\n", " array([[[1.2402402]]], dtype=float32),\n", " array([[[1.239]]], dtype=float32),\n", " array([[[1.2377613]]], dtype=float32),\n", " array([[[1.2365221]]], dtype=float32),\n", " array([[[1.2352988]]], dtype=float32),\n", " array([[[1.2340796]]], dtype=float32),\n", " array([[[1.2328496]]], dtype=float32),\n", " array([[[1.2316185]]], dtype=float32),\n", " array([[[1.2303916]]], dtype=float32),\n", " array([[[1.2291664]]], dtype=float32),\n", " array([[[1.2279422]]], dtype=float32),\n", " array([[[1.2267293]]], dtype=float32),\n", " array([[[1.2255194]]], dtype=float32),\n", " array([[[1.2243121]]], dtype=float32),\n", " array([[[1.223104]]], dtype=float32),\n", " array([[[1.221898]]], dtype=float32),\n", " array([[[1.220692]]], dtype=float32),\n", " array([[[1.2194864]]], dtype=float32),\n", " array([[[1.2182887]]], dtype=float32),\n", " array([[[1.2170899]]], dtype=float32),\n", " array([[[1.215891]]], dtype=float32),\n", " array([[[1.2146913]]], dtype=float32),\n", " array([[[1.2134917]]], dtype=float32),\n", " array([[[1.2122788]]], dtype=float32),\n", " array([[[1.211078]]], dtype=float32),\n", " array([[[1.2098831]]], dtype=float32),\n", " array([[[1.2086855]]], dtype=float32),\n", " array([[[1.2074825]]], dtype=float32),\n", " array([[[1.2062756]]], dtype=float32),\n", " array([[[1.2050577]]], dtype=float32),\n", " array([[[1.2038466]]], dtype=float32),\n", " array([[[1.2026336]]], dtype=float32),\n", " array([[[1.201427]]], dtype=float32),\n", " array([[[1.2002189]]], dtype=float32),\n", " array([[[1.1990134]]], dtype=float32),\n", " array([[[1.197809]]], dtype=float32),\n", " array([[[1.1966054]]], dtype=float32),\n", " array([[[1.1954057]]], dtype=float32),\n", " array([[[1.1942087]]], dtype=float32),\n", " array([[[1.1930114]]], dtype=float32),\n", " array([[[1.1918113]]], dtype=float32),\n", " array([[[1.190607]]], dtype=float32),\n", " array([[[1.1894008]]], dtype=float32),\n", " array([[[1.1881926]]], dtype=float32),\n", " array([[[1.1869885]]], dtype=float32),\n", " array([[[1.1857878]]], dtype=float32),\n", " array([[[1.1845922]]], dtype=float32),\n", " array([[[1.1833982]]], dtype=float32),\n", " array([[[1.1822051]]], dtype=float32),\n", " array([[[1.1810207]]], dtype=float32),\n", " array([[[1.179842]]], dtype=float32),\n", " array([[[1.178653]]], dtype=float32),\n", " array([[[1.1774693]]], dtype=float32),\n", " array([[[1.1762929]]], dtype=float32),\n", " array([[[1.1751094]]], dtype=float32),\n", " array([[[1.1739267]]], dtype=float32),\n", " array([[[1.1727498]]], dtype=float32),\n", " array([[[1.1715744]]], dtype=float32),\n", " array([[[1.1704013]]], dtype=float32),\n", " array([[[1.1692263]]], dtype=float32),\n", " array([[[1.1680505]]], dtype=float32),\n", " array([[[1.1668729]]], dtype=float32),\n", " array([[[1.1656915]]], dtype=float32),\n", " array([[[1.1645228]]], dtype=float32),\n", " array([[[1.1633562]]], dtype=float32),\n", " array([[[1.162192]]], dtype=float32),\n", " array([[[1.1610341]]], dtype=float32),\n", " array([[[1.1598762]]], dtype=float32),\n", " array([[[1.1587218]]], dtype=float32),\n", " array([[[1.1575752]]], dtype=float32),\n", " array([[[1.1564318]]], dtype=float32),\n", " array([[[1.1552829]]], dtype=float32),\n", " array([[[1.1541296]]], dtype=float32),\n", " array([[[1.1529925]]], dtype=float32),\n", " array([[[1.1518486]]], dtype=float32),\n", " array([[[1.150708]]], dtype=float32),\n", " array([[[1.1495737]]], dtype=float32),\n", " array([[[1.1484416]]], dtype=float32),\n", " array([[[1.1473128]]], dtype=float32),\n", " array([[[1.1461834]]], dtype=float32),\n", " array([[[1.1450512]]], dtype=float32),\n", " array([[[1.14392]]], dtype=float32),\n", " array([[[1.1427876]]], dtype=float32),\n", " array([[[1.1416541]]], dtype=float32),\n", " array([[[1.1405251]]], dtype=float32),\n", " array([[[1.1393929]]], dtype=float32),\n", " array([[[1.1382619]]], dtype=float32),\n", " array([[[1.1371318]]], dtype=float32),\n", " array([[[1.1360084]]], dtype=float32),\n", " array([[[1.1348764]]], dtype=float32),\n", " array([[[1.13374]]], dtype=float32),\n", " array([[[1.1326076]]], dtype=float32),\n", " array([[[1.1314675]]], dtype=float32),\n", " array([[[1.1303337]]], dtype=float32),\n", " array([[[1.1291977]]], dtype=float32),\n", " array([[[1.1280602]]], dtype=float32),\n", " array([[[1.1269315]]], dtype=float32),\n", " array([[[1.125806]]], dtype=float32),\n", " array([[[1.1246831]]], dtype=float32),\n", " array([[[1.1235552]]], dtype=float32),\n", " array([[[1.1224195]]], dtype=float32),\n", " array([[[1.121293]]], dtype=float32),\n", " array([[[1.1201667]]], dtype=float32),\n", " array([[[1.1190311]]], dtype=float32),\n", " array([[[1.1178986]]], dtype=float32),\n", " array([[[1.1167703]]], dtype=float32),\n", " array([[[1.1156447]]], dtype=float32),\n", " array([[[1.1145234]]], dtype=float32),\n", " array([[[1.1134056]]], dtype=float32),\n", " array([[[1.1122894]]], dtype=float32),\n", " array([[[1.1111698]]], dtype=float32),\n", " array([[[1.1100578]]], dtype=float32),\n", " array([[[1.1089429]]], dtype=float32),\n", " array([[[1.1078281]]], dtype=float32),\n", " array([[[1.1067177]]], dtype=float32),\n", " array([[[1.105607]]], dtype=float32),\n", " array([[[1.1044995]]], dtype=float32),\n", " array([[[1.1033937]]], dtype=float32),\n", " array([[[1.1022813]]], dtype=float32),\n", " array([[[1.101175]]], dtype=float32),\n", " array([[[1.100072]]], dtype=float32),\n", " array([[[1.0989705]]], dtype=float32),\n", " array([[[1.09787]]], dtype=float32),\n", " array([[[1.0967644]]], dtype=float32),\n", " array([[[1.0956646]]], dtype=float32),\n", " array([[[1.0945646]]], dtype=float32),\n", " array([[[1.0934709]]], dtype=float32),\n", " array([[[1.0923792]]], dtype=float32),\n", " array([[[1.0912739]]], dtype=float32),\n", " array([[[1.0901752]]], dtype=float32),\n", " array([[[1.0890846]]], dtype=float32),\n", " array([[[1.0879909]]], dtype=float32),\n", " array([[[1.0869026]]], dtype=float32),\n", " array([[[1.085813]]], dtype=float32),\n", " array([[[1.0847273]]], dtype=float32),\n", " array([[[1.083642]]], dtype=float32),\n", " array([[[1.0825577]]], dtype=float32),\n", " array([[[1.0814714]]], dtype=float32),\n", " array([[[1.0803818]]], dtype=float32),\n", " array([[[1.0792935]]], dtype=float32),\n", " array([[[1.0782086]]], dtype=float32),\n", " array([[[1.077128]]], dtype=float32),\n", " array([[[1.076051]]], dtype=float32),\n", " array([[[1.0749768]]], dtype=float32),\n", " array([[[1.0739008]]], dtype=float32),\n", " array([[[1.0728284]]], dtype=float32),\n", " array([[[1.0717567]]], dtype=float32),\n", " array([[[1.0706822]]], dtype=float32),\n", " array([[[1.0696145]]], dtype=float32),\n", " array([[[1.0685458]]], dtype=float32),\n", " array([[[1.0674801]]], dtype=float32),\n", " array([[[1.0664152]]], dtype=float32),\n", " array([[[1.0653456]]], dtype=float32),\n", " array([[[1.0642809]]], dtype=float32),\n", " array([[[1.0632174]]], dtype=float32),\n", " array([[[1.0621449]]], dtype=float32),\n", " array([[[1.0610808]]], dtype=float32),\n", " array([[[1.0600215]]], dtype=float32),\n", " array([[[1.0589641]]], dtype=float32),\n", " array([[[1.0579076]]], dtype=float32),\n", " array([[[1.056849]]], dtype=float32),\n", " array([[[1.0557933]]], dtype=float32),\n", " array([[[1.0547335]]], dtype=float32),\n", " array([[[1.0536795]]], dtype=float32),\n", " array([[[1.052621]]], dtype=float32),\n", " array([[[1.0515645]]], dtype=float32),\n", " array([[[1.0505136]]], dtype=float32),\n", " array([[[1.0494663]]], dtype=float32),\n", " array([[[1.0484185]]], dtype=float32),\n", " array([[[1.0473726]]], dtype=float32),\n", " array([[[1.0463277]]], dtype=float32),\n", " array([[[1.0452754]]], dtype=float32),\n", " array([[[1.0442339]]], dtype=float32),\n", " array([[[1.0431925]]], dtype=float32),\n", " array([[[1.0421488]]], dtype=float32),\n", " array([[[1.0411083]]], dtype=float32),\n", " array([[[1.0400652]]], dtype=float32),\n", " array([[[1.0390263]]], dtype=float32),\n", " array([[[1.0379914]]], dtype=float32),\n", " array([[[1.036955]]], dtype=float32),\n", " array([[[1.035918]]], dtype=float32),\n", " array([[[1.0348793]]], dtype=float32),\n", " array([[[1.0338428]]], dtype=float32),\n", " array([[[1.0328108]]], dtype=float32),\n", " array([[[1.0317804]]], dtype=float32),\n", " array([[[1.0307503]]], dtype=float32),\n", " array([[[1.0297221]]], dtype=float32),\n", " array([[[1.0286949]]], dtype=float32),\n", " array([[[1.0276692]]], dtype=float32),\n", " array([[[1.0266424]]], dtype=float32),\n", " array([[[1.025617]]], dtype=float32),\n", " array([[[1.0245935]]], dtype=float32),\n", " array([[[1.023571]]], dtype=float32),\n", " array([[[1.0225477]]], dtype=float32),\n", " array([[[1.0215256]]], dtype=float32),\n", " array([[[1.0205063]]], dtype=float32),\n", " array([[[1.0194863]]], dtype=float32),\n", " array([[[1.0184672]]], dtype=float32),\n", " array([[[1.0174494]]], dtype=float32),\n", " array([[[1.0164313]]], dtype=float32),\n", " array([[[1.0154179]]], dtype=float32),\n", " array([[[1.014403]]], dtype=float32),\n", " array([[[1.0133876]]], dtype=float32),\n", " array([[[1.012376]]], dtype=float32),\n", " array([[[1.0113665]]], dtype=float32),\n", " array([[[1.0103574]]], dtype=float32),\n", " array([[[1.009349]]], dtype=float32),\n", " array([[[1.0083407]]], dtype=float32),\n", " array([[[1.0073277]]], dtype=float32),\n", " array([[[1.0063177]]], dtype=float32),\n", " array([[[1.0053138]]], dtype=float32),\n", " array([[[1.004312]]], dtype=float32),\n", " array([[[1.0033094]]], dtype=float32),\n", " array([[[1.0023073]]], dtype=float32),\n", " array([[[1.0013069]]], dtype=float32),\n", " array([[[1.0003057]]], dtype=float32),\n", " array([[[0.9993069]]], dtype=float32),\n", " array([[[0.9983047]]], dtype=float32),\n", " array([[[0.9973078]]], dtype=float32),\n", " array([[[0.9963064]]], dtype=float32),\n", " array([[[0.9953128]]], dtype=float32),\n", " array([[[0.9943164]]], dtype=float32),\n", " array([[[0.99332124]]], dtype=float32),\n", " array([[[0.99233013]]], dtype=float32),\n", " array([[[0.9913407]]], dtype=float32),\n", " array([[[0.99035347]]], dtype=float32),\n", " array([[[0.98936623]]], dtype=float32),\n", " array([[[0.98837835]]], dtype=float32),\n", " array([[[0.98738956]]], dtype=float32),\n", " array([[[0.98640394]]], dtype=float32),\n", " array([[[0.9854166]]], dtype=float32),\n", " array([[[0.9844333]]], dtype=float32),\n", " array([[[0.9834489]]], dtype=float32),\n", " array([[[0.98245966]]], dtype=float32),\n", " array([[[0.9814797]]], dtype=float32),\n", " array([[[0.98050106]]], dtype=float32),\n", " array([[[0.9795218]]], dtype=float32),\n", " array([[[0.97854495]]], dtype=float32),\n", " array([[[0.977566]]], dtype=float32),\n", " array([[[0.976587]]], dtype=float32),\n", " array([[[0.9756114]]], dtype=float32),\n", " array([[[0.97463757]]], dtype=float32),\n", " array([[[0.9736647]]], dtype=float32),\n", " array([[[0.972694]]], dtype=float32),\n", " array([[[0.97172415]]], dtype=float32),\n", " array([[[0.9707535]]], dtype=float32),\n", " array([[[0.96978486]]], dtype=float32),\n", " array([[[0.9688183]]], dtype=float32),\n", " array([[[0.96785116]]], dtype=float32),\n", " array([[[0.9668832]]], dtype=float32),\n", " array([[[0.9659168]]], dtype=float32),\n", " array([[[0.9649522]]], dtype=float32),\n", " array([[[0.963988]]], dtype=float32),\n", " array([[[0.9630251]]], dtype=float32),\n", " array([[[0.96206474]]], dtype=float32),\n", " array([[[0.96110433]]], dtype=float32),\n", " array([[[0.9601409]]], dtype=float32),\n", " array([[[0.95917577]]], dtype=float32),\n", " array([[[0.95821553]]], dtype=float32),\n", " array([[[0.9572568]]], dtype=float32),\n", " array([[[0.9563019]]], dtype=float32),\n", " array([[[0.9553434]]], dtype=float32),\n", " array([[[0.95438826]]], dtype=float32),\n", " array([[[0.95343226]]], dtype=float32),\n", " array([[[0.9524765]]], dtype=float32),\n", " array([[[0.9515256]]], dtype=float32),\n", " array([[[0.9505773]]], dtype=float32),\n", " array([[[0.94962883]]], dtype=float32),\n", " array([[[0.9486766]]], dtype=float32),\n", " array([[[0.9477276]]], dtype=float32),\n", " array([[[0.94677997]]], dtype=float32),\n", " array([[[0.94583523]]], dtype=float32),\n", " array([[[0.9448923]]], dtype=float32),\n", " array([[[0.9439499]]], dtype=float32),\n", " array([[[0.9430058]]], dtype=float32),\n", " array([[[0.9420611]]], dtype=float32),\n", " array([[[0.9411204]]], dtype=float32),\n", " array([[[0.94018203]]], dtype=float32),\n", " array([[[0.93924254]]], dtype=float32),\n", " array([[[0.93830323]]], dtype=float32),\n", " array([[[0.9373658]]], dtype=float32),\n", " array([[[0.93642926]]], dtype=float32),\n", " array([[[0.9354959]]], dtype=float32),\n", " array([[[0.93456]]], dtype=float32),\n", " array([[[0.9336257]]], dtype=float32),\n", " array([[[0.9326926]]], dtype=float32),\n", " array([[[0.931761]]], dtype=float32),\n", " array([[[0.9308281]]], dtype=float32),\n", " array([[[0.92989534]]], dtype=float32),\n", " array([[[0.9289666]]], dtype=float32),\n", " array([[[0.9280401]]], dtype=float32),\n", " array([[[0.9271147]]], dtype=float32),\n", " array([[[0.9261888]]], dtype=float32),\n", " array([[[0.92526376]]], dtype=float32),\n", " array([[[0.9243406]]], dtype=float32),\n", " array([[[0.9234169]]], dtype=float32),\n", " array([[[0.9224933]]], dtype=float32),\n", " array([[[0.921574]]], dtype=float32),\n", " array([[[0.92065513]]], dtype=float32),\n", " array([[[0.91973424]]], dtype=float32),\n", " array([[[0.91881394]]], dtype=float32),\n", " array([[[0.91789114]]], dtype=float32),\n", " array([[[0.9169698]]], dtype=float32),\n", " array([[[0.9160542]]], dtype=float32),\n", " array([[[0.9151343]]], dtype=float32),\n", " array([[[0.91421366]]], dtype=float32),\n", " array([[[0.9133012]]], dtype=float32),\n", " array([[[0.9123889]]], dtype=float32),\n", " array([[[0.9114798]]], dtype=float32),\n", " array([[[0.9105729]]], dtype=float32),\n", " array([[[0.9096676]]], dtype=float32),\n", " array([[[0.9087631]]], dtype=float32),\n", " array([[[0.9078585]]], dtype=float32),\n", " array([[[0.90695363]]], dtype=float32),\n", " array([[[0.9060471]]], dtype=float32),\n", " array([[[0.9051384]]], dtype=float32),\n", " array([[[0.9042276]]], dtype=float32),\n", " array([[[0.9033194]]], dtype=float32),\n", " array([[[0.902418]]], dtype=float32),\n", " array([[[0.90151596]]], dtype=float32),\n", " array([[[0.9006148]]], dtype=float32),\n", " array([[[0.8997174]]], dtype=float32),\n", " array([[[0.8988177]]], dtype=float32),\n", " array([[[0.8979216]]], dtype=float32),\n", " array([[[0.897029]]], dtype=float32),\n", " array([[[0.8961339]]], dtype=float32),\n", " array([[[0.8952393]]], dtype=float32),\n", " array([[[0.8943479]]], dtype=float32),\n", " array([[[0.8934551]]], dtype=float32),\n", " array([[[0.89256227]]], dtype=float32),\n", " array([[[0.89167035]]], dtype=float32),\n", " array([[[0.8907801]]], dtype=float32),\n", " array([[[0.8898902]]], dtype=float32),\n", " array([[[0.88900155]]], dtype=float32),\n", " array([[[0.88811636]]], dtype=float32),\n", " array([[[0.88723207]]], dtype=float32),\n", " array([[[0.8863468]]], dtype=float32),\n", " array([[[0.8854627]]], dtype=float32),\n", " array([[[0.88457817]]], dtype=float32),\n", " array([[[0.88369244]]], dtype=float32),\n", " array([[[0.8828095]]], dtype=float32),\n", " array([[[0.881929]]], dtype=float32),\n", " array([[[0.8810496]]], dtype=float32),\n", " array([[[0.8801712]]], dtype=float32),\n", " array([[[0.8792936]]], dtype=float32),\n", " array([[[0.8784152]]], dtype=float32),\n", " array([[[0.877536]]], dtype=float32),\n", " array([[[0.87665683]]], dtype=float32),\n", " array([[[0.87578136]]], dtype=float32),\n", " array([[[0.87490827]]], dtype=float32),\n", " array([[[0.8740353]]], dtype=float32),\n", " array([[[0.87316227]]], dtype=float32),\n", " array([[[0.8722893]]], dtype=float32),\n", " array([[[0.8714153]]], dtype=float32),\n", " array([[[0.8705437]]], dtype=float32),\n", " array([[[0.8696735]]], dtype=float32),\n", " array([[[0.8688036]]], dtype=float32),\n", " array([[[0.867937]]], dtype=float32),\n", " array([[[0.86707157]]], dtype=float32),\n", " array([[[0.86620367]]], dtype=float32),\n", " array([[[0.8653372]]], dtype=float32),\n", " array([[[0.8644714]]], dtype=float32),\n", " array([[[0.86360496]]], dtype=float32),\n", " array([[[0.8627414]]], dtype=float32),\n", " array([[[0.861879]]], dtype=float32),\n", " array([[[0.8610171]]], dtype=float32),\n", " array([[[0.8601578]]], dtype=float32),\n", " array([[[0.8593015]]], dtype=float32),\n", " array([[[0.85844564]]], dtype=float32),\n", " array([[[0.8575876]]], dtype=float32),\n", " array([[[0.85673]]], dtype=float32),\n", " array([[[0.8558732]]], dtype=float32),\n", " array([[[0.85501736]]], dtype=float32),\n", " array([[[0.8541631]]], dtype=float32),\n", " array([[[0.8533105]]], dtype=float32),\n", " array([[[0.85245913]]], dtype=float32),\n", " array([[[0.85160834]]], dtype=float32),\n", " array([[[0.85075885]]], dtype=float32),\n", " array([[[0.84991014]]], dtype=float32),\n", " array([[[0.8490621]]], dtype=float32),\n", " array([[[0.8482149]]], dtype=float32),\n", " array([[[0.84736913]]], dtype=float32),\n", " array([[[0.84652394]]], dtype=float32),\n", " array([[[0.8456782]]], dtype=float32),\n", " array([[[0.84483284]]], dtype=float32),\n", " array([[[0.84398913]]], dtype=float32),\n", " array([[[0.8431424]]], dtype=float32),\n", " array([[[0.84228957]]], dtype=float32),\n", " array([[[0.8414443]]], dtype=float32),\n", " array([[[0.8406052]]], dtype=float32),\n", " array([[[0.8397668]]], dtype=float32),\n", " array([[[0.83892727]]], dtype=float32),\n", " array([[[0.8380877]]], dtype=float32),\n", " array([[[0.8372508]]], dtype=float32),\n", " array([[[0.8364148]]], dtype=float32),\n", " array([[[0.8355804]]], dtype=float32),\n", " array([[[0.83474624]]], dtype=float32),\n", " array([[[0.83391035]]], dtype=float32),\n", " array([[[0.83307475]]], dtype=float32),\n", " array([[[0.8322414]]], dtype=float32),\n", " array([[[0.83140993]]], dtype=float32),\n", " array([[[0.8305805]]], dtype=float32),\n", " array([[[0.82975304]]], dtype=float32),\n", " array([[[0.8289262]]], dtype=float32),\n", " array([[[0.8281006]]], dtype=float32),\n", " array([[[0.8272754]]], dtype=float32),\n", " array([[[0.8264486]]], dtype=float32),\n", " array([[[0.82562083]]], dtype=float32),\n", " array([[[0.82479507]]], dtype=float32),\n", " array([[[0.8239707]]], dtype=float32),\n", " array([[[0.82314795]]], dtype=float32),\n", " array([[[0.82232773]]], dtype=float32),\n", " array([[[0.82150775]]], dtype=float32),\n", " array([[[0.82068735]]], dtype=float32),\n", " array([[[0.8198682]]], dtype=float32),\n", " array([[[0.81905067]]], dtype=float32),\n", " array([[[0.8182336]]], dtype=float32),\n", " array([[[0.8174171]]], dtype=float32),\n", " array([[[0.8166016]]], dtype=float32),\n", " array([[[0.8157869]]], dtype=float32),\n", " array([[[0.8149721]]], dtype=float32),\n", " array([[[0.814157]]], dtype=float32),\n", " array([[[0.8133414]]], dtype=float32),\n", " array([[[0.8125268]]], dtype=float32),\n", " array([[[0.81171423]]], dtype=float32),\n", " array([[[0.81090355]]], dtype=float32),\n", " array([[[0.8100937]]], dtype=float32),\n", " array([[[0.8092845]]], dtype=float32),\n", " array([[[0.8084758]]], dtype=float32),\n", " array([[[0.8076685]]], dtype=float32),\n", " array([[[0.80686194]]], dtype=float32),\n", " array([[[0.80605376]]], dtype=float32),\n", " array([[[0.8052466]]], dtype=float32),\n", " array([[[0.80444217]]], dtype=float32),\n", " array([[[0.8036392]]], dtype=float32),\n", " array([[[0.8028356]]], dtype=float32),\n", " array([[[0.80203056]]], dtype=float32),\n", " array([[[0.8012261]]], dtype=float32),\n", " array([[[0.80042535]]], dtype=float32),\n", " array([[[0.7996267]]], dtype=float32),\n", " array([[[0.7988291]]], dtype=float32),\n", " array([[[0.79803073]]], dtype=float32),\n", " array([[[0.79723024]]], dtype=float32),\n", " array([[[0.7964298]]], dtype=float32),\n", " array([[[0.7956321]]], dtype=float32),\n", " array([[[0.7948366]]], dtype=float32),\n", " array([[[0.7940424]]], dtype=float32),\n", " array([[[0.7932502]]], dtype=float32),\n", " array([[[0.7924593]]], dtype=float32),\n", " array([[[0.79166734]]], dtype=float32),\n", " array([[[0.7908751]]], dtype=float32),\n", " array([[[0.790083]]], dtype=float32),\n", " array([[[0.7892926]]], dtype=float32),\n", " array([[[0.7885065]]], dtype=float32),\n", " array([[[0.78772277]]], dtype=float32),\n", " array([[[0.78693926]]], dtype=float32),\n", " array([[[0.78615266]]], dtype=float32),\n", " array([[[0.785361]]], dtype=float32),\n", " array([[[0.784572]]], dtype=float32),\n", " array([[[0.7837901]]], dtype=float32),\n", " array([[[0.78300995]]], dtype=float32),\n", " array([[[0.782229]]], dtype=float32),\n", " array([[[0.7814491]]], dtype=float32),\n", " array([[[0.7806714]]], dtype=float32),\n", " array([[[0.7798939]]], dtype=float32),\n", " array([[[0.7791158]]], dtype=float32),\n", " array([[[0.7783377]]], dtype=float32),\n", " array([[[0.7775601]]], dtype=float32),\n", " array([[[0.7767833]]], dtype=float32),\n", " array([[[0.77600616]]], dtype=float32),\n", " array([[[0.77522576]]], dtype=float32),\n", " array([[[0.7744434]]], dtype=float32),\n", " array([[[0.7736673]]], dtype=float32),\n", " array([[[0.7728975]]], dtype=float32),\n", " array([[[0.7721281]]], dtype=float32),\n", " array([[[0.7713584]]], dtype=float32),\n", " array([[[0.77058995]]], dtype=float32),\n", " array([[[0.7698235]]], dtype=float32),\n", " array([[[0.76905835]]], dtype=float32),\n", " array([[[0.7682923]]], dtype=float32),\n", " array([[[0.76752377]]], dtype=float32),\n", " array([[[0.76675355]]], dtype=float32),\n", " array([[[0.7659839]]], dtype=float32),\n", " array([[[0.7652158]]], dtype=float32),\n", " array([[[0.76445013]]], dtype=float32),\n", " array([[[0.76368594]]], dtype=float32),\n", " array([[[0.76292074]]], dtype=float32),\n", " array([[[0.76215804]]], dtype=float32),\n", " array([[[0.7614002]]], dtype=float32),\n", " array([[[0.7606441]]], dtype=float32),\n", " array([[[0.7598884]]], dtype=float32),\n", " array([[[0.7591333]]], dtype=float32),\n", " array([[[0.7583773]]], dtype=float32),\n", " array([[[0.7576191]]], dtype=float32),\n", " array([[[0.75685835]]], dtype=float32),\n", " array([[[0.7560993]]], dtype=float32),\n", " array([[[0.7553441]]], dtype=float32),\n", " array([[[0.7545908]]], dtype=float32),\n", " array([[[0.7538388]]], dtype=float32),\n", " array([[[0.7530867]]], dtype=float32),\n", " array([[[0.75233436]]], dtype=float32),\n", " array([[[0.7515839]]], dtype=float32),\n", " array([[[0.7508357]]], dtype=float32),\n", " array([[[0.7500884]]], dtype=float32),\n", " array([[[0.74933964]]], dtype=float32),\n", " array([[[0.7485877]]], dtype=float32),\n", " array([[[0.74783635]]], dtype=float32),\n", " array([[[0.74708956]]], dtype=float32),\n", " array([[[0.7463435]]], dtype=float32),\n", " array([[[0.7455954]]], dtype=float32),\n", " array([[[0.7448493]]], dtype=float32),\n", " array([[[0.7441072]]], dtype=float32),\n", " array([[[0.74336773]]], dtype=float32),\n", " array([[[0.74262935]]], dtype=float32),\n", " array([[[0.74189025]]], dtype=float32),\n", " array([[[0.74114984]]], dtype=float32),\n", " array([[[0.74040806]]], dtype=float32),\n", " array([[[0.7396675]]], dtype=float32),\n", " array([[[0.7389297]]], dtype=float32),\n", " array([[[0.7381918]]], dtype=float32),\n", " array([[[0.7374516]]], dtype=float32),\n", " array([[[0.73671293]]], dtype=float32),\n", " array([[[0.73597866]]], dtype=float32),\n", " array([[[0.73524624]]], dtype=float32),\n", " array([[[0.7345133]]], dtype=float32),\n", " array([[[0.7337806]]], dtype=float32),\n", " array([[[0.73304915]]], dtype=float32),\n", " array([[[0.7323186]]], dtype=float32),\n", " array([[[0.7315889]]], dtype=float32),\n", " array([[[0.7308603]]], dtype=float32),\n", " array([[[0.73013175]]], dtype=float32),\n", " array([[[0.7294025]]], dtype=float32),\n", " array([[[0.72867304]]], dtype=float32),\n", " array([[[0.72794455]]], dtype=float32),\n", " array([[[0.72721785]]], dtype=float32),\n", " array([[[0.726492]]], dtype=float32),\n", " array([[[0.7257652]]], dtype=float32),\n", " array([[[0.7250375]]], dtype=float32),\n", " array([[[0.7243103]]], dtype=float32),\n", " array([[[0.7235842]]], dtype=float32),\n", " array([[[0.7228604]]], dtype=float32),\n", " array([[[0.72214]]], dtype=float32),\n", " array([[[0.7214218]]], dtype=float32),\n", " array([[[0.7207037]]], dtype=float32),\n", " array([[[0.7199849]]], dtype=float32),\n", " array([[[0.7192661]]], dtype=float32),\n", " array([[[0.7185486]]], dtype=float32),\n", " array([[[0.7178323]]], dtype=float32),\n", " array([[[0.7171153]]], dtype=float32),\n", " array([[[0.7163955]]], dtype=float32),\n", " array([[[0.7156745]]], dtype=float32),\n", " array([[[0.7149559]]], dtype=float32),\n", " array([[[0.7142406]]], dtype=float32),\n", " array([[[0.71352726]]], dtype=float32),\n", " array([[[0.7128148]]], dtype=float32),\n", " array([[[0.7121036]]], dtype=float32),\n", " array([[[0.7113942]]], dtype=float32),\n", " array([[[0.7106859]]], dtype=float32),\n", " array([[[0.70997745]]], dtype=float32),\n", " array([[[0.7092679]]], dtype=float32),\n", " array([[[0.7085571]]], dtype=float32),\n", " array([[[0.70784706]]], dtype=float32),\n", " array([[[0.7071392]]], dtype=float32),\n", " array([[[0.70643353]]], dtype=float32),\n", " array([[[0.70572865]]], dtype=float32),\n", " array([[[0.7050222]]], dtype=float32),\n", " array([[[0.704315]]], dtype=float32),\n", " array([[[0.7036106]]], dtype=float32),\n", " array([[[0.7029095]]], dtype=float32),\n", " array([[[0.7022088]]], dtype=float32),\n", " array([[[0.7015075]]], dtype=float32),\n", " array([[[0.70080656]]], dtype=float32),\n", " array([[[0.700106]]], dtype=float32),\n", " array([[[0.6994046]]], dtype=float32),\n", " array([[[0.69870174]]], dtype=float32),\n", " array([[[0.6979994]]], dtype=float32),\n", " array([[[0.69730127]]], dtype=float32),\n", " array([[[0.69660705]]], dtype=float32),\n", " array([[[0.6959145]]], dtype=float32),\n", " array([[[0.6952227]]], dtype=float32),\n", " array([[[0.6945303]]], dtype=float32),\n", " array([[[0.6938373]]], dtype=float32),\n", " array([[[0.69314456]]], dtype=float32),\n", " array([[[0.6924525]]], dtype=float32),\n", " array([[[0.69176114]]], dtype=float32),\n", " array([[[0.6910708]]], dtype=float32),\n", " array([[[0.69038117]]], dtype=float32),\n", " array([[[0.68969154]]], dtype=float32),\n", " array([[[0.68900055]]], dtype=float32),\n", " array([[[0.68830824]]], dtype=float32),\n", " array([[[0.6876166]]], dtype=float32),\n", " array([[[0.68692666]]], dtype=float32),\n", " array([[[0.68623847]]], dtype=float32),\n", " array([[[0.68555254]]], dtype=float32),\n", " array([[[0.68486774]]], dtype=float32),\n", " array([[[0.684183]]], dtype=float32),\n", " array([[[0.68349975]]], dtype=float32),\n", " array([[[0.68281806]]], dtype=float32),\n", " array([[[0.68213654]]], dtype=float32),\n", " array([[[0.6814544]]], dtype=float32),\n", " array([[[0.6807734]]], dtype=float32),\n", " array([[[0.6800952]]], dtype=float32),\n", " array([[[0.67941916]]], dtype=float32),\n", " array([[[0.67874295]]], dtype=float32),\n", " array([[[0.6780654]]], dtype=float32),\n", " array([[[0.67738736]]], dtype=float32),\n", " array([[[0.6767106]]], dtype=float32),\n", " array([[[0.6760361]]], dtype=float32),\n", " array([[[0.6753624]]], dtype=float32),\n", " array([[[0.6746883]]], dtype=float32),\n", " array([[[0.6740151]]], dtype=float32),\n", " array([[[0.6733437]]], dtype=float32),\n", " array([[[0.6726738]]], dtype=float32),\n", " array([[[0.67200434]]], dtype=float32),\n", " array([[[0.67133224]]], dtype=float32),\n", " array([[[0.67065537]]], dtype=float32),\n", " array([[[0.6699774]]], dtype=float32),\n", " array([[[0.6693048]]], dtype=float32),\n", " array([[[0.66863793]]], dtype=float32),\n", " array([[[0.6679733]]], dtype=float32),\n", " array([[[0.6673082]]], dtype=float32),\n", " array([[[0.666643]]], dtype=float32),\n", " array([[[0.6659788]]], dtype=float32),\n", " array([[[0.6653146]]], dtype=float32),\n", " array([[[0.6646487]]], dtype=float32),\n", " array([[[0.6639812]]], dtype=float32),\n", " array([[[0.6633147]]], dtype=float32),\n", " array([[[0.6626512]]], dtype=float32),\n", " array([[[0.6619906]]], dtype=float32),\n", " array([[[0.66133237]]], dtype=float32),\n", " array([[[0.66067547]]], dtype=float32),\n", " array([[[0.6600186]]], dtype=float32),\n", " array([[[0.659361]]], dtype=float32),\n", " array([[[0.6587025]]], dtype=float32),\n", " array([[[0.6580442]]], dtype=float32),\n", " array([[[0.6573873]]], dtype=float32),\n", " array([[[0.6567322]]], dtype=float32),\n", " array([[[0.6560776]]], dtype=float32),\n", " array([[[0.65542203]]], dtype=float32),\n", " array([[[0.65476507]]], dtype=float32),\n", " array([[[0.65410954]]], dtype=float32),\n", " array([[[0.65345716]]], dtype=float32),\n", " array([[[0.6528071]]], dtype=float32),\n", " array([[[0.652158]]], dtype=float32),\n", " array([[[0.65150875]]], dtype=float32),\n", " array([[[0.6508589]]], dtype=float32),\n", " array([[[0.65020907]]], dtype=float32),\n", " array([[[0.64956045]]], dtype=float32),\n", " array([[[0.6489127]]], dtype=float32),\n", " array([[[0.64826506]]], dtype=float32),\n", " array([[[0.64761794]]], dtype=float32),\n", " array([[[0.6469725]]], dtype=float32),\n", " array([[[0.64632875]]], dtype=float32),\n", " array([[[0.64568573]]], dtype=float32),\n", " array([[[0.6450428]]], dtype=float32),\n", " array([[[0.64439934]]], dtype=float32),\n", " array([[[0.6437556]]], dtype=float32),\n", " array([[[0.6431124]]], dtype=float32),\n", " array([[[0.6424704]]], dtype=float32),\n", " array([[[0.6418295]]], dtype=float32),\n", " array([[[0.6411892]]], dtype=float32),\n", " array([[[0.6405496]]], dtype=float32),\n", " array([[[0.6399103]]], dtype=float32),\n", " array([[[0.6392699]]], dtype=float32),\n", " array([[[0.6386267]]], dtype=float32),\n", " array([[[0.6379811]]], dtype=float32),\n", " array([[[0.63733566]]], dtype=float32),\n", " array([[[0.6366924]]], dtype=float32),\n", " array([[[0.6360531]]], dtype=float32),\n", " array([[[0.6354181]]], dtype=float32),\n", " array([[[0.6347859]]], dtype=float32),\n", " array([[[0.6341539]]], dtype=float32),\n", " array([[[0.63352203]]], dtype=float32),\n", " array([[[0.63289136]]], dtype=float32),\n", " array([[[0.63226193]]], dtype=float32),\n", " array([[[0.6316336]]], dtype=float32),\n", " array([[[0.63100535]]], dtype=float32),\n", " array([[[0.6303756]]], dtype=float32),\n", " array([[[0.6297434]]], dtype=float32),\n", " array([[[0.6291104]]], dtype=float32),\n", " array([[[0.6284789]]], dtype=float32),\n", " array([[[0.6278486]]], dtype=float32),\n", " array([[[0.6272183]]], dtype=float32),\n", " array([[[0.6265875]]], dtype=float32),\n", " array([[[0.6259569]]], dtype=float32),\n", " array([[[0.6253278]]], dtype=float32),\n", " array([[[0.62470204]]], dtype=float32),\n", " array([[[0.6240794]]], dtype=float32),\n", " array([[[0.62345785]]], dtype=float32),\n", " array([[[0.62283504]]], dtype=float32),\n", " array([[[0.6222111]]], dtype=float32),\n", " array([[[0.62158775]]], dtype=float32),\n", " array([[[0.6209658]]], dtype=float32),\n", " array([[[0.6203462]]], dtype=float32),\n", " array([[[0.61972934]]], dtype=float32),\n", " array([[[0.61911446]]], dtype=float32),\n", " array([[[0.6185003]]], dtype=float32),\n", " array([[[0.61788607]]], dtype=float32),\n", " array([[[0.61727065]]], dtype=float32),\n", " array([[[0.61665255]]], dtype=float32),\n", " array([[[0.61603296]]], dtype=float32),\n", " array([[[0.6154146]]], dtype=float32),\n", " array([[[0.61479884]]], dtype=float32),\n", " array([[[0.61418355]]], dtype=float32),\n", " array([[[0.61356664]]], dtype=float32),\n", " array([[[0.6129498]]], dtype=float32),\n", " array([[[0.6123364]]], dtype=float32),\n", " array([[[0.61172646]]], dtype=float32),\n", " array([[[0.6111177]]], dtype=float32),\n", " array([[[0.6105091]]], dtype=float32),\n", " array([[[0.60990137]]], dtype=float32),\n", " array([[[0.6092958]]], dtype=float32),\n", " array([[[0.6086928]]], dtype=float32),\n", " array([[[0.6080918]]], dtype=float32),\n", " array([[[0.60749084]]], dtype=float32),\n", " array([[[0.6068884]]], dtype=float32),\n", " array([[[0.6062841]]], dtype=float32),\n", " array([[[0.6056797]]], dtype=float32),\n", " array([[[0.6050767]]], dtype=float32),\n", " array([[[0.60447466]]], dtype=float32),\n", " array([[[0.6038714]]], dtype=float32),\n", " array([[[0.60326433]]], dtype=float32),\n", " array([[[0.60265356]]], dtype=float32),\n", " array([[[0.60204273]]], dtype=float32),\n", " array([[[0.60143614]]], dtype=float32),\n", " array([[[0.6008345]]], dtype=float32),\n", " array([[[0.60023457]]], dtype=float32),\n", " array([[[0.5996337]]], dtype=float32),\n", " array([[[0.599033]]], dtype=float32),\n", " array([[[0.5984356]]], dtype=float32),\n", " array([[[0.59784245]]], dtype=float32),\n", " array([[[0.59725153]]], dtype=float32),\n", " array([[[0.59665865]]], dtype=float32),\n", " array([[[0.5960615]]], dtype=float32),\n", " array([[[0.59546167]]], dtype=float32),\n", " array([[[0.5948632]]], dtype=float32),\n", " array([[[0.594269]]], dtype=float32),\n", " array([[[0.5936787]]], dtype=float32),\n", " array([[[0.5930896]]], dtype=float32),\n", " array([[[0.59249973]]], dtype=float32),\n", " array([[[0.5919075]]], dtype=float32),\n", " array([[[0.5913137]]], dtype=float32),\n", " array([[[0.59072065]]], dtype=float32),\n", " array([[[0.59013075]]], dtype=float32),\n", " array([[[0.5895443]]], dtype=float32),\n", " array([[[0.5889603]]], dtype=float32),\n", " array([[[0.58837813]]], dtype=float32),\n", " array([[[0.5877974]]], dtype=float32),\n", " array([[[0.58721656]]], dtype=float32),\n", " array([[[0.5866334]]], dtype=float32),\n", " array([[[0.5860466]]], dtype=float32),\n", " array([[[0.585457]]], dtype=float32),\n", " array([[[0.5848677]]], dtype=float32),\n", " array([[[0.58428085]]], dtype=float32),\n", " array([[[0.58369654]]], dtype=float32),\n", " array([[[0.5831127]]], dtype=float32),\n", " array([[[0.5825283]]], dtype=float32),\n", " array([[[0.5819432]]], dtype=float32),\n", " array([[[0.5813586]]], dtype=float32),\n", " array([[[0.58077466]]], dtype=float32),\n", " array([[[0.58019024]]], dtype=float32),\n", " array([[[0.5796043]]], dtype=float32),\n", " array([[[0.5790163]]], dtype=float32),\n", " array([[[0.57842803]]], dtype=float32),\n", " array([[[0.5778427]]], dtype=float32),\n", " array([[[0.57726306]]], dtype=float32),\n", " array([[[0.576689]]], dtype=float32),\n", " array([[[0.5761191]]], dtype=float32),\n", " array([[[0.5755524]]], dtype=float32),\n", " array([[[0.57498807]]], dtype=float32),\n", " array([[[0.5744259]]], dtype=float32),\n", " array([[[0.57386494]]], dtype=float32),\n", " array([[[0.5733041]]], dtype=float32),\n", " array([[[0.5727428]]], dtype=float32),\n", " array([[[0.5721801]]], dtype=float32),\n", " array([[[0.57161534]]], dtype=float32),\n", " array([[[0.5710484]]], dtype=float32),\n", " array([[[0.5704798]]], dtype=float32),\n", " array([[[0.56991047]]], dtype=float32),\n", " array([[[0.56934047]]], dtype=float32),\n", " array([[[0.5687696]]], dtype=float32),\n", " array([[[0.5681981]]], dtype=float32),\n", " array([[[0.56762683]]], dtype=float32),\n", " array([[[0.5670567]]], dtype=float32),\n", " array([[[0.5664877]]], dtype=float32),\n", " array([[[0.5659199]]], dtype=float32),\n", " array([[[0.5653535]]], dtype=float32),\n", " array([[[0.5647888]]], dtype=float32),\n", " array([[[0.564226]]], dtype=float32),\n", " array([[[0.56366444]]], dtype=float32),\n", " array([[[0.5631039]]], dtype=float32),\n", " array([[[0.5625443]]], dtype=float32),\n", " array([[[0.56198525]]], dtype=float32),\n", " array([[[0.56142664]]], dtype=float32),\n", " array([[[0.56086826]]], dtype=float32),\n", " array([[[0.5603106]]], dtype=float32),\n", " array([[[0.55975354]]], dtype=float32),\n", " array([[[0.5591967]]], dtype=float32),\n", " array([[[0.5586398]]], dtype=float32),\n", " array([[[0.5580827]]], dtype=float32),\n", " array([[[0.55752575]]], dtype=float32),\n", " array([[[0.5569695]]], dtype=float32),\n", " array([[[0.5564143]]], dtype=float32),\n", " array([[[0.5558598]]], dtype=float32),\n", " array([[[0.55530554]]], dtype=float32),\n", " array([[[0.5547511]]], dtype=float32),\n", " array([[[0.5541963]]], dtype=float32),\n", " array([[[0.5536418]]], dtype=float32),\n", " array([[[0.5530882]]], dtype=float32),\n", " array([[[0.5525357]]], dtype=float32),\n", " array([[[0.55198437]]], dtype=float32),\n", " array([[[0.551434]]], dtype=float32),\n", " array([[[0.5508845]]], dtype=float32),\n", " array([[[0.55033535]]], dtype=float32),\n", " array([[[0.54978585]]], dtype=float32),\n", " array([[[0.5492346]]], dtype=float32),\n", " array([[[0.5486804]]], dtype=float32),\n", " array([[[0.5481225]]], dtype=float32),\n", " array([[[0.54756254]]], dtype=float32),\n", " array([[[0.54700327]]], dtype=float32),\n", " array([[[0.546448]]], dtype=float32),\n", " array([[[0.5458983]]], dtype=float32),\n", " array([[[0.54535425]]], dtype=float32),\n", " array([[[0.54481465]]], dtype=float32),\n", " array([[[0.54427785]]], dtype=float32),\n", " array([[[0.5437421]]], dtype=float32),\n", " array([[[0.5432061]]], dtype=float32),\n", " array([[[0.5426688]]], dtype=float32),\n", " array([[[0.5421301]]], dtype=float32),\n", " array([[[0.54159033]]], dtype=float32),\n", " array([[[0.5410501]]], dtype=float32),\n", " array([[[0.54050976]]], dtype=float32),\n", " array([[[0.53996974]]], dtype=float32),\n", " array([[[0.5394303]]], dtype=float32),\n", " array([[[0.5388919]]], dtype=float32),\n", " array([[[0.53835446]]], dtype=float32),\n", " array([[[0.53781706]]], dtype=float32),\n", " array([[[0.5372788]]], dtype=float32),\n", " array([[[0.53673923]]], dtype=float32),\n", " array([[[0.53619933]]], dtype=float32),\n", " array([[[0.53566045]]], dtype=float32),\n", " array([[[0.5351239]]], dtype=float32),\n", " array([[[0.53458995]]], dtype=float32),\n", " array([[[0.53405845]]], dtype=float32),\n", " array([[[0.5335291]]], dtype=float32),\n", " array([[[0.5330012]]], dtype=float32),\n", " array([[[0.53247374]]], dtype=float32),\n", " array([[[0.53194577]]], dtype=float32),\n", " array([[[0.5314165]]], dtype=float32),\n", " array([[[0.53088516]]], dtype=float32),\n", " array([[[0.5303523]]], dtype=float32),\n", " array([[[0.5298189]]], dtype=float32),\n", " array([[[0.5292868]]], dtype=float32),\n", " array([[[0.5287573]]], dtype=float32),\n", " array([[[0.5282305]]], dtype=float32),\n", " array([[[0.5277055]]], dtype=float32),\n", " array([[[0.52718055]]], dtype=float32),\n", " array([[[0.5266545]]], dtype=float32),\n", " array([[[0.5261268]]], dtype=float32),\n", " array([[[0.52559763]]], dtype=float32),\n", " array([[[0.52506816]]], dtype=float32),\n", " array([[[0.52453965]]], dtype=float32),\n", " array([[[0.5240129]]], dtype=float32),\n", " array([[[0.52348816]]], dtype=float32),\n", " array([[[0.52296436]]], dtype=float32),\n", " array([[[0.5224401]]], dtype=float32),\n", " array([[[0.521914]]], dtype=float32),\n", " array([[[0.521386]]], dtype=float32),\n", " array([[[0.5208575]]], dtype=float32),\n", " array([[[0.5203307]]], dtype=float32),\n", " array([[[0.5198077]]], dtype=float32),\n", " array([[[0.51928914]]], dtype=float32),\n", " ...],\n", " 'covar_module.mixture_scales': [array([[[0.00155212]]], dtype=float32),\n", " array([[[0.00148847]]], dtype=float32),\n", " array([[[0.00142766]]], dtype=float32),\n", " array([[[0.00136703]]], dtype=float32),\n", " array([[[0.00131087]]], dtype=float32),\n", " array([[[0.00125889]]], dtype=float32),\n", " array([[[0.00121174]]], dtype=float32),\n", " array([[[0.00116328]]], dtype=float32),\n", " array([[[0.00111723]]], dtype=float32),\n", " array([[[0.00107112]]], dtype=float32),\n", " array([[[0.00103181]]], dtype=float32),\n", " array([[[0.00099283]]], dtype=float32),\n", " array([[[0.00095359]]], dtype=float32),\n", " array([[[0.00091442]]], dtype=float32),\n", " array([[[0.00087738]]], dtype=float32),\n", " array([[[0.00084221]]], dtype=float32),\n", " array([[[0.00080906]]], dtype=float32),\n", " array([[[0.00077662]]], dtype=float32),\n", " array([[[0.00074534]]], dtype=float32),\n", " array([[[0.00071365]]], dtype=float32),\n", " array([[[0.00068499]]], dtype=float32),\n", " array([[[0.00066115]]], dtype=float32),\n", " array([[[0.00063648]]], dtype=float32),\n", " array([[[0.00061464]]], dtype=float32),\n", " array([[[0.00059508]]], dtype=float32),\n", " array([[[0.00057613]]], dtype=float32),\n", " array([[[0.00055894]]], dtype=float32),\n", " array([[[0.00054079]]], dtype=float32),\n", " array([[[0.00052403]]], dtype=float32),\n", " array([[[0.00050698]]], dtype=float32),\n", " array([[[0.00048862]]], dtype=float32),\n", " array([[[0.00047098]]], dtype=float32),\n", " array([[[0.00045334]]], dtype=float32),\n", " array([[[0.00043577]]], dtype=float32),\n", " array([[[0.00041815]]], dtype=float32),\n", " array([[[0.00040135]]], dtype=float32),\n", " array([[[0.00038587]]], dtype=float32),\n", " array([[[0.00037214]]], dtype=float32),\n", " array([[[0.00036025]]], dtype=float32),\n", " array([[[0.00034794]]], dtype=float32),\n", " array([[[0.00033686]]], dtype=float32),\n", " array([[[0.00032689]]], dtype=float32),\n", " array([[[0.00031795]]], dtype=float32),\n", " array([[[0.00030994]]], dtype=float32),\n", " array([[[0.0003028]]], dtype=float32),\n", " array([[[0.00029568]]], dtype=float32),\n", " array([[[0.00028842]]], dtype=float32),\n", " array([[[0.00028073]]], dtype=float32),\n", " array([[[0.0002735]]], dtype=float32),\n", " array([[[0.00026727]]], dtype=float32),\n", " array([[[0.00026076]]], dtype=float32),\n", " array([[[0.00025421]]], dtype=float32),\n", " array([[[0.00024745]]], dtype=float32),\n", " array([[[0.00024089]]], dtype=float32),\n", " array([[[0.0002349]]], dtype=float32),\n", " array([[[0.00022974]]], dtype=float32),\n", " array([[[0.00022469]]], dtype=float32),\n", " array([[[0.00021963]]], dtype=float32),\n", " array([[[0.0002149]]], dtype=float32),\n", " array([[[0.00021051]]], dtype=float32),\n", " array([[[0.00020633]]], dtype=float32),\n", " array([[[0.00020214]]], dtype=float32),\n", " array([[[0.00019866]]], dtype=float32),\n", " array([[[0.00019601]]], dtype=float32),\n", " array([[[0.00019312]]], dtype=float32),\n", " array([[[0.00019092]]], dtype=float32),\n", " array([[[0.0001893]]], dtype=float32),\n", " array([[[0.00018765]]], dtype=float32),\n", " array([[[0.00018632]]], dtype=float32),\n", " array([[[0.00018483]]], dtype=float32),\n", " array([[[0.00018399]]], dtype=float32),\n", " array([[[0.00018283]]], dtype=float32),\n", " array([[[0.00018145]]], dtype=float32),\n", " array([[[0.00017987]]], dtype=float32),\n", " array([[[0.00017862]]], dtype=float32),\n", " array([[[0.00017798]]], dtype=float32),\n", " array([[[0.00017714]]], dtype=float32),\n", " array([[[0.00017646]]], dtype=float32),\n", " array([[[0.00017672]]], dtype=float32),\n", " array([[[0.00017773]]], dtype=float32),\n", " array([[[0.00017875]]], dtype=float32),\n", " array([[[0.00018017]]], dtype=float32),\n", " array([[[0.0001813]]], dtype=float32),\n", " array([[[0.00018204]]], dtype=float32),\n", " array([[[0.00018228]]], dtype=float32),\n", " array([[[0.0001824]]], dtype=float32),\n", " array([[[0.00018233]]], dtype=float32),\n", " array([[[0.0001822]]], dtype=float32),\n", " array([[[0.00018236]]], dtype=float32),\n", " array([[[0.00018265]]], dtype=float32),\n", " array([[[0.00018352]]], dtype=float32),\n", " array([[[0.00018447]]], dtype=float32),\n", " array([[[0.00018483]]], dtype=float32),\n", " array([[[0.00018484]]], dtype=float32),\n", " array([[[0.0001845]]], dtype=float32),\n", " array([[[0.000184]]], dtype=float32),\n", " array([[[0.00018314]]], dtype=float32),\n", " array([[[0.00018309]]], dtype=float32),\n", " array([[[0.00018289]]], dtype=float32),\n", " array([[[0.00018235]]], dtype=float32),\n", " array([[[0.00018202]]], dtype=float32),\n", " array([[[0.00018208]]], dtype=float32),\n", " array([[[0.00018221]]], dtype=float32),\n", " array([[[0.00018238]]], dtype=float32),\n", " array([[[0.00018246]]], dtype=float32),\n", " array([[[0.00018235]]], dtype=float32),\n", " array([[[0.00018261]]], dtype=float32),\n", " array([[[0.00018278]]], dtype=float32),\n", " array([[[0.00018278]]], dtype=float32),\n", " array([[[0.00018243]]], dtype=float32),\n", " array([[[0.00018196]]], dtype=float32),\n", " array([[[0.00018148]]], dtype=float32),\n", " array([[[0.000181]]], dtype=float32),\n", " array([[[0.00018064]]], dtype=float32),\n", " array([[[0.00017982]]], dtype=float32),\n", " array([[[0.00017936]]], dtype=float32),\n", " array([[[0.0001789]]], dtype=float32),\n", " array([[[0.00017822]]], dtype=float32),\n", " array([[[0.00017769]]], dtype=float32),\n", " array([[[0.00017719]]], dtype=float32),\n", " array([[[0.00017656]]], dtype=float32),\n", " array([[[0.00017588]]], dtype=float32),\n", " array([[[0.00017525]]], dtype=float32),\n", " array([[[0.0001748]]], dtype=float32),\n", " array([[[0.00017425]]], dtype=float32),\n", " array([[[0.00017386]]], dtype=float32),\n", " array([[[0.00017365]]], dtype=float32),\n", " array([[[0.00017342]]], dtype=float32),\n", " array([[[0.00017327]]], dtype=float32),\n", " array([[[0.000173]]], dtype=float32),\n", " array([[[0.00017267]]], dtype=float32),\n", " array([[[0.00017229]]], dtype=float32),\n", " array([[[0.00017223]]], dtype=float32),\n", " array([[[0.00017237]]], dtype=float32),\n", " array([[[0.00017252]]], dtype=float32),\n", " array([[[0.00017266]]], dtype=float32),\n", " array([[[0.00017264]]], dtype=float32),\n", " array([[[0.00017249]]], dtype=float32),\n", " array([[[0.00017226]]], dtype=float32),\n", " array([[[0.00017194]]], dtype=float32),\n", " array([[[0.00017152]]], dtype=float32),\n", " array([[[0.00017096]]], dtype=float32),\n", " array([[[0.00017053]]], dtype=float32),\n", " array([[[0.00017017]]], dtype=float32),\n", " array([[[0.00016977]]], dtype=float32),\n", " array([[[0.00016926]]], dtype=float32),\n", " array([[[0.00016882]]], dtype=float32),\n", " array([[[0.00016816]]], dtype=float32),\n", " array([[[0.00016734]]], dtype=float32),\n", " array([[[0.0001664]]], dtype=float32),\n", " array([[[0.00016546]]], dtype=float32),\n", " array([[[0.0001649]]], dtype=float32),\n", " array([[[0.00016512]]], dtype=float32),\n", " array([[[0.00016553]]], dtype=float32),\n", " array([[[0.00016591]]], dtype=float32),\n", " array([[[0.00016636]]], dtype=float32),\n", " array([[[0.0001671]]], dtype=float32),\n", " array([[[0.00016827]]], dtype=float32),\n", " array([[[0.00016956]]], dtype=float32),\n", " array([[[0.00017073]]], dtype=float32),\n", " array([[[0.00017175]]], dtype=float32),\n", " array([[[0.00017243]]], dtype=float32),\n", " array([[[0.0001729]]], dtype=float32),\n", " array([[[0.00017299]]], dtype=float32),\n", " array([[[0.00017303]]], dtype=float32),\n", " array([[[0.00017303]]], dtype=float32),\n", " array([[[0.00017299]]], dtype=float32),\n", " array([[[0.00017303]]], dtype=float32),\n", " array([[[0.00017369]]], dtype=float32),\n", " array([[[0.00017429]]], dtype=float32),\n", " array([[[0.00017477]]], dtype=float32),\n", " array([[[0.00017517]]], dtype=float32),\n", " array([[[0.00017542]]], dtype=float32),\n", " array([[[0.00017609]]], dtype=float32),\n", " array([[[0.00017705]]], dtype=float32),\n", " array([[[0.00017781]]], dtype=float32),\n", " array([[[0.00017809]]], dtype=float32),\n", " array([[[0.00017801]]], dtype=float32),\n", " array([[[0.00017811]]], dtype=float32),\n", " array([[[0.0001783]]], dtype=float32),\n", " array([[[0.00017888]]], dtype=float32),\n", " array([[[0.0001795]]], dtype=float32),\n", " array([[[0.00018029]]], dtype=float32),\n", " array([[[0.00018102]]], dtype=float32),\n", " array([[[0.00018173]]], dtype=float32),\n", " array([[[0.00018234]]], dtype=float32),\n", " array([[[0.00018301]]], dtype=float32),\n", " array([[[0.00018342]]], dtype=float32),\n", " array([[[0.00018389]]], dtype=float32),\n", " array([[[0.00018441]]], dtype=float32),\n", " array([[[0.00018476]]], dtype=float32),\n", " array([[[0.00018512]]], dtype=float32),\n", " array([[[0.00018562]]], dtype=float32),\n", " array([[[0.00018625]]], dtype=float32),\n", " array([[[0.00018692]]], dtype=float32),\n", " array([[[0.00018744]]], dtype=float32),\n", " array([[[0.00018794]]], dtype=float32),\n", " array([[[0.00018852]]], dtype=float32),\n", " array([[[0.000189]]], dtype=float32),\n", " array([[[0.00018951]]], dtype=float32),\n", " array([[[0.00019058]]], dtype=float32),\n", " array([[[0.00019261]]], dtype=float32),\n", " array([[[0.0001946]]], dtype=float32),\n", " array([[[0.00019743]]], dtype=float32),\n", " array([[[0.0002005]]], dtype=float32),\n", " array([[[0.00020318]]], dtype=float32),\n", " array([[[0.00020525]]], dtype=float32),\n", " array([[[0.00020698]]], dtype=float32),\n", " array([[[0.00020869]]], dtype=float32),\n", " array([[[0.00021033]]], dtype=float32),\n", " array([[[0.00021177]]], dtype=float32),\n", " array([[[0.00021325]]], dtype=float32),\n", " array([[[0.000215]]], dtype=float32),\n", " array([[[0.00021653]]], dtype=float32),\n", " array([[[0.00021793]]], dtype=float32),\n", " array([[[0.00021888]]], dtype=float32),\n", " array([[[0.0002195]]], dtype=float32),\n", " array([[[0.00021973]]], dtype=float32),\n", " array([[[0.00022044]]], dtype=float32),\n", " array([[[0.00022095]]], dtype=float32),\n", " array([[[0.00022214]]], dtype=float32),\n", " array([[[0.00022418]]], dtype=float32),\n", " array([[[0.00022614]]], dtype=float32),\n", " array([[[0.00022896]]], dtype=float32),\n", " array([[[0.0002324]]], dtype=float32),\n", " array([[[0.00023587]]], dtype=float32),\n", " array([[[0.00023893]]], dtype=float32),\n", " array([[[0.00024173]]], dtype=float32),\n", " array([[[0.00024524]]], dtype=float32),\n", " array([[[0.00024938]]], dtype=float32),\n", " array([[[0.00025305]]], dtype=float32),\n", " array([[[0.00025682]]], dtype=float32),\n", " array([[[0.0002609]]], dtype=float32),\n", " array([[[0.00026507]]], dtype=float32),\n", " array([[[0.00026909]]], dtype=float32),\n", " array([[[0.0002727]]], dtype=float32),\n", " array([[[0.00027594]]], dtype=float32),\n", " array([[[0.00027855]]], dtype=float32),\n", " array([[[0.00028134]]], dtype=float32),\n", " array([[[0.00028326]]], dtype=float32),\n", " array([[[0.00028369]]], dtype=float32),\n", " array([[[0.00028439]]], dtype=float32),\n", " array([[[0.00028512]]], dtype=float32),\n", " array([[[0.00028625]]], dtype=float32),\n", " array([[[0.00028761]]], dtype=float32),\n", " array([[[0.00028901]]], dtype=float32),\n", " array([[[0.00028975]]], dtype=float32),\n", " array([[[0.00029088]]], dtype=float32),\n", " array([[[0.0002919]]], dtype=float32),\n", " array([[[0.00029262]]], dtype=float32),\n", " array([[[0.00029308]]], dtype=float32),\n", " array([[[0.00029402]]], dtype=float32),\n", " array([[[0.00029541]]], dtype=float32),\n", " array([[[0.00029744]]], dtype=float32),\n", " array([[[0.00029978]]], dtype=float32),\n", " array([[[0.00030164]]], dtype=float32),\n", " array([[[0.00030395]]], dtype=float32),\n", " array([[[0.00030733]]], dtype=float32),\n", " array([[[0.00031106]]], dtype=float32),\n", " array([[[0.00031449]]], dtype=float32),\n", " array([[[0.00031785]]], dtype=float32),\n", " array([[[0.00032113]]], dtype=float32),\n", " array([[[0.00032347]]], dtype=float32),\n", " array([[[0.00032553]]], dtype=float32),\n", " array([[[0.00032767]]], dtype=float32),\n", " array([[[0.00032867]]], dtype=float32),\n", " array([[[0.00032965]]], dtype=float32),\n", " array([[[0.00033233]]], dtype=float32),\n", " array([[[0.00033524]]], dtype=float32),\n", " array([[[0.00033868]]], dtype=float32),\n", " array([[[0.00034209]]], dtype=float32),\n", " array([[[0.00034483]]], dtype=float32),\n", " array([[[0.00034773]]], dtype=float32),\n", " array([[[0.00035012]]], dtype=float32),\n", " array([[[0.0003524]]], dtype=float32),\n", " array([[[0.00035473]]], dtype=float32),\n", " array([[[0.00035702]]], dtype=float32),\n", " array([[[0.00035945]]], dtype=float32),\n", " array([[[0.00036127]]], dtype=float32),\n", " array([[[0.00036292]]], dtype=float32),\n", " array([[[0.00036492]]], dtype=float32),\n", " array([[[0.00036619]]], dtype=float32),\n", " array([[[0.00036794]]], dtype=float32),\n", " array([[[0.0003694]]], dtype=float32),\n", " array([[[0.0003728]]], dtype=float32),\n", " array([[[0.00037574]]], dtype=float32),\n", " array([[[0.00037835]]], dtype=float32),\n", " array([[[0.00038074]]], dtype=float32),\n", " array([[[0.00038302]]], dtype=float32),\n", " array([[[0.00038468]]], dtype=float32),\n", " array([[[0.00038603]]], dtype=float32),\n", " array([[[0.00038745]]], dtype=float32),\n", " array([[[0.00039002]]], dtype=float32),\n", " array([[[0.00039364]]], dtype=float32),\n", " array([[[0.0003954]]], dtype=float32),\n", " array([[[0.00039771]]], dtype=float32),\n", " array([[[0.0004004]]], dtype=float32),\n", " array([[[0.00040247]]], dtype=float32),\n", " array([[[0.00040429]]], dtype=float32),\n", " array([[[0.00040628]]], dtype=float32),\n", " array([[[0.0004088]]], dtype=float32),\n", " array([[[0.0004112]]], dtype=float32),\n", " array([[[0.00041305]]], dtype=float32),\n", " array([[[0.00041532]]], dtype=float32),\n", " array([[[0.0004171]]], dtype=float32),\n", " array([[[0.0004194]]], dtype=float32),\n", " array([[[0.00042152]]], dtype=float32),\n", " array([[[0.00042382]]], dtype=float32),\n", " array([[[0.00042559]]], dtype=float32),\n", " array([[[0.00042753]]], dtype=float32),\n", " array([[[0.00042995]]], dtype=float32),\n", " array([[[0.00043155]]], dtype=float32),\n", " array([[[0.00043406]]], dtype=float32),\n", " array([[[0.00043517]]], dtype=float32),\n", " array([[[0.00043691]]], dtype=float32),\n", " array([[[0.00043834]]], dtype=float32),\n", " array([[[0.00043988]]], dtype=float32),\n", " array([[[0.00044107]]], dtype=float32),\n", " array([[[0.00044221]]], dtype=float32),\n", " array([[[0.00044329]]], dtype=float32),\n", " array([[[0.00044444]]], dtype=float32),\n", " array([[[0.00044611]]], dtype=float32),\n", " array([[[0.00044613]]], dtype=float32),\n", " array([[[0.00044707]]], dtype=float32),\n", " array([[[0.00044767]]], dtype=float32),\n", " array([[[0.00044872]]], dtype=float32),\n", " array([[[0.00044954]]], dtype=float32),\n", " array([[[0.00044986]]], dtype=float32),\n", " array([[[0.0004508]]], dtype=float32),\n", " array([[[0.00045177]]], dtype=float32),\n", " array([[[0.0004521]]], dtype=float32),\n", " array([[[0.00045328]]], dtype=float32),\n", " array([[[0.00045404]]], dtype=float32),\n", " array([[[0.00045491]]], dtype=float32),\n", " array([[[0.00045566]]], dtype=float32),\n", " array([[[0.00045586]]], dtype=float32),\n", " array([[[0.00045867]]], dtype=float32),\n", " array([[[0.00045807]]], dtype=float32),\n", " array([[[0.00045919]]], dtype=float32),\n", " array([[[0.00046048]]], dtype=float32),\n", " array([[[0.00046198]]], dtype=float32),\n", " array([[[0.0004625]]], dtype=float32),\n", " array([[[0.00046388]]], dtype=float32),\n", " array([[[0.00046404]]], dtype=float32),\n", " array([[[0.00046548]]], dtype=float32),\n", " array([[[0.00046628]]], dtype=float32),\n", " array([[[0.00046754]]], dtype=float32),\n", " array([[[0.00046885]]], dtype=float32),\n", " array([[[0.00046996]]], dtype=float32),\n", " array([[[0.00047271]]], dtype=float32),\n", " array([[[0.00047374]]], dtype=float32),\n", " array([[[0.00047542]]], dtype=float32),\n", " array([[[0.00047705]]], dtype=float32),\n", " array([[[0.00047858]]], dtype=float32),\n", " array([[[0.00047999]]], dtype=float32),\n", " array([[[0.00048163]]], dtype=float32),\n", " array([[[0.00048229]]], dtype=float32),\n", " array([[[0.0004836]]], dtype=float32),\n", " array([[[0.00048507]]], dtype=float32),\n", " array([[[0.00048593]]], dtype=float32),\n", " array([[[0.0004876]]], dtype=float32),\n", " array([[[0.00048886]]], dtype=float32),\n", " array([[[0.00048981]]], dtype=float32),\n", " array([[[0.00049145]]], dtype=float32),\n", " array([[[0.00049305]]], dtype=float32),\n", " array([[[0.00049421]]], dtype=float32),\n", " array([[[0.00049581]]], dtype=float32),\n", " array([[[0.0004968]]], dtype=float32),\n", " array([[[0.00049782]]], dtype=float32),\n", " array([[[0.000499]]], dtype=float32),\n", " array([[[0.00049977]]], dtype=float32),\n", " array([[[0.00050062]]], dtype=float32),\n", " array([[[0.00050154]]], dtype=float32),\n", " array([[[0.00050225]]], dtype=float32),\n", " array([[[0.00050252]]], dtype=float32),\n", " array([[[0.00050304]]], dtype=float32),\n", " array([[[0.00050349]]], dtype=float32),\n", " array([[[0.00050433]]], dtype=float32),\n", " array([[[0.00050436]]], dtype=float32),\n", " array([[[0.00050472]]], dtype=float32),\n", " array([[[0.00050528]]], dtype=float32),\n", " array([[[0.00050514]]], dtype=float32),\n", " array([[[0.00050545]]], dtype=float32),\n", " array([[[0.00050567]]], dtype=float32),\n", " array([[[0.00050721]]], dtype=float32),\n", " array([[[0.00050692]]], dtype=float32),\n", " array([[[0.00050733]]], dtype=float32),\n", " array([[[0.00050771]]], dtype=float32),\n", " array([[[0.00050852]]], dtype=float32),\n", " array([[[0.00050986]]], dtype=float32),\n", " array([[[0.0005107]]], dtype=float32),\n", " array([[[0.00051206]]], dtype=float32),\n", " array([[[0.00051312]]], dtype=float32),\n", " array([[[0.00051434]]], dtype=float32),\n", " array([[[0.00051549]]], dtype=float32),\n", " array([[[0.00051725]]], dtype=float32),\n", " array([[[0.00051914]]], dtype=float32),\n", " array([[[0.00051985]]], dtype=float32),\n", " array([[[0.00052134]]], dtype=float32),\n", " array([[[0.00052208]]], dtype=float32),\n", " array([[[0.00052347]]], dtype=float32),\n", " array([[[0.00052445]]], dtype=float32),\n", " array([[[0.00052658]]], dtype=float32),\n", " array([[[0.000527]]], dtype=float32),\n", " array([[[0.00052826]]], dtype=float32),\n", " array([[[0.00052923]]], dtype=float32),\n", " array([[[0.00053121]]], dtype=float32),\n", " array([[[0.00053193]]], dtype=float32),\n", " array([[[0.0005334]]], dtype=float32),\n", " array([[[0.00053407]]], dtype=float32),\n", " array([[[0.00053537]]], dtype=float32),\n", " array([[[0.0005369]]], dtype=float32),\n", " array([[[0.00053774]]], dtype=float32),\n", " array([[[0.00053842]]], dtype=float32),\n", " array([[[0.00054003]]], dtype=float32),\n", " array([[[0.00054095]]], dtype=float32),\n", " array([[[0.00054248]]], dtype=float32),\n", " array([[[0.00054302]]], dtype=float32),\n", " array([[[0.00054448]]], dtype=float32),\n", " array([[[0.00054583]]], dtype=float32),\n", " array([[[0.00054738]]], dtype=float32),\n", " array([[[0.00054828]]], dtype=float32),\n", " array([[[0.0005493]]], dtype=float32),\n", " array([[[0.00055085]]], dtype=float32),\n", " array([[[0.00055147]]], dtype=float32),\n", " array([[[0.00055213]]], dtype=float32),\n", " array([[[0.00055362]]], dtype=float32),\n", " array([[[0.00055474]]], dtype=float32),\n", " array([[[0.00055599]]], dtype=float32),\n", " array([[[0.00055784]]], dtype=float32),\n", " array([[[0.00055809]]], dtype=float32),\n", " array([[[0.00055931]]], dtype=float32),\n", " array([[[0.00056285]]], dtype=float32),\n", " array([[[0.00056295]]], dtype=float32),\n", " array([[[0.00056458]]], dtype=float32),\n", " array([[[0.00056673]]], dtype=float32),\n", " array([[[0.00056797]]], dtype=float32),\n", " array([[[0.00056979]]], dtype=float32),\n", " array([[[0.00057116]]], dtype=float32),\n", " array([[[0.00057255]]], dtype=float32),\n", " array([[[0.0005739]]], dtype=float32),\n", " array([[[0.00057503]]], dtype=float32),\n", " array([[[0.00057662]]], dtype=float32),\n", " array([[[0.0005781]]], dtype=float32),\n", " array([[[0.00057967]]], dtype=float32),\n", " array([[[0.00057971]]], dtype=float32),\n", " array([[[0.00058046]]], dtype=float32),\n", " array([[[0.00058314]]], dtype=float32),\n", " array([[[0.00058361]]], dtype=float32),\n", " array([[[0.00058593]]], dtype=float32),\n", " array([[[0.0005879]]], dtype=float32),\n", " array([[[0.00058844]]], dtype=float32),\n", " array([[[0.00059008]]], dtype=float32),\n", " array([[[0.00059246]]], dtype=float32),\n", " array([[[0.00059302]]], dtype=float32),\n", " array([[[0.00059428]]], dtype=float32),\n", " array([[[0.00059629]]], dtype=float32),\n", " array([[[0.00059711]]], dtype=float32),\n", " array([[[0.00059858]]], dtype=float32),\n", " array([[[0.00059929]]], dtype=float32),\n", " array([[[0.00060122]]], dtype=float32),\n", " array([[[0.00060176]]], dtype=float32),\n", " array([[[0.00060266]]], dtype=float32),\n", " array([[[0.000604]]], dtype=float32),\n", " array([[[0.00060523]]], dtype=float32),\n", " array([[[0.00060578]]], dtype=float32),\n", " array([[[0.00060746]]], dtype=float32),\n", " array([[[0.00060818]]], dtype=float32),\n", " array([[[0.00060859]]], dtype=float32),\n", " array([[[0.00060928]]], dtype=float32),\n", " array([[[0.00061004]]], dtype=float32),\n", " array([[[0.00061056]]], dtype=float32),\n", " array([[[0.00061133]]], dtype=float32),\n", " array([[[0.00061221]]], dtype=float32),\n", " array([[[0.00061345]]], dtype=float32),\n", " array([[[0.00061366]]], dtype=float32),\n", " array([[[0.00061362]]], dtype=float32),\n", " array([[[0.00061419]]], dtype=float32),\n", " array([[[0.00061496]]], dtype=float32),\n", " array([[[0.00061554]]], dtype=float32),\n", " array([[[0.00061634]]], dtype=float32),\n", " array([[[0.00061691]]], dtype=float32),\n", " array([[[0.00061669]]], dtype=float32),\n", " array([[[0.00061752]]], dtype=float32),\n", " array([[[0.00061766]]], dtype=float32),\n", " array([[[0.00061735]]], dtype=float32),\n", " array([[[0.00061835]]], dtype=float32),\n", " array([[[0.00061947]]], dtype=float32),\n", " array([[[0.00061935]]], dtype=float32),\n", " array([[[0.00062079]]], dtype=float32),\n", " array([[[0.00062123]]], dtype=float32),\n", " array([[[0.00062188]]], dtype=float32),\n", " array([[[0.00062278]]], dtype=float32),\n", " array([[[0.00062381]]], dtype=float32),\n", " array([[[0.00062378]]], dtype=float32),\n", " array([[[0.00062444]]], dtype=float32),\n", " array([[[0.0006257]]], dtype=float32),\n", " array([[[0.00062705]]], dtype=float32),\n", " array([[[0.00062731]]], dtype=float32),\n", " array([[[0.00062821]]], dtype=float32),\n", " array([[[0.00062845]]], dtype=float32),\n", " array([[[0.00062898]]], dtype=float32),\n", " array([[[0.00062925]]], dtype=float32),\n", " array([[[0.0006299]]], dtype=float32),\n", " array([[[0.0006303]]], dtype=float32),\n", " array([[[0.00063059]]], dtype=float32),\n", " array([[[0.00063097]]], dtype=float32),\n", " array([[[0.00063134]]], dtype=float32),\n", " array([[[0.00063143]]], dtype=float32),\n", " array([[[0.00063148]]], dtype=float32),\n", " array([[[0.00063176]]], dtype=float32),\n", " array([[[0.00063225]]], dtype=float32),\n", " array([[[0.00063209]]], dtype=float32),\n", " array([[[0.00063216]]], dtype=float32),\n", " array([[[0.0006348]]], dtype=float32),\n", " array([[[0.00063547]]], dtype=float32),\n", " array([[[0.00063344]]], dtype=float32),\n", " array([[[0.00063414]]], dtype=float32),\n", " array([[[0.00063534]]], dtype=float32),\n", " array([[[0.00063696]]], dtype=float32),\n", " array([[[0.00063738]]], dtype=float32),\n", " array([[[0.00063788]]], dtype=float32),\n", " array([[[0.00063881]]], dtype=float32),\n", " array([[[0.00063921]]], dtype=float32),\n", " array([[[0.00064077]]], dtype=float32),\n", " array([[[0.00064208]]], dtype=float32),\n", " array([[[0.00064283]]], dtype=float32),\n", " array([[[0.00064317]]], dtype=float32),\n", " array([[[0.00064401]]], dtype=float32),\n", " array([[[0.00064437]]], dtype=float32),\n", " array([[[0.00064515]]], dtype=float32),\n", " array([[[0.00064591]]], dtype=float32),\n", " array([[[0.00064637]]], dtype=float32),\n", " array([[[0.00064738]]], dtype=float32),\n", " array([[[0.00064874]]], dtype=float32),\n", " array([[[0.00064928]]], dtype=float32),\n", " array([[[0.00064939]]], dtype=float32),\n", " array([[[0.00064998]]], dtype=float32),\n", " array([[[0.0006499]]], dtype=float32),\n", " array([[[0.00065017]]], dtype=float32),\n", " array([[[0.00065114]]], dtype=float32),\n", " array([[[0.00065164]]], dtype=float32),\n", " array([[[0.00065166]]], dtype=float32),\n", " array([[[0.00065188]]], dtype=float32),\n", " array([[[0.00065234]]], dtype=float32),\n", " array([[[0.00065249]]], dtype=float32),\n", " array([[[0.0006525]]], dtype=float32),\n", " array([[[0.00065273]]], dtype=float32),\n", " array([[[0.0006531]]], dtype=float32),\n", " array([[[0.00065367]]], dtype=float32),\n", " array([[[0.00065398]]], dtype=float32),\n", " array([[[0.00065391]]], dtype=float32),\n", " array([[[0.00065355]]], dtype=float32),\n", " array([[[0.00065368]]], dtype=float32),\n", " array([[[0.00065374]]], dtype=float32),\n", " array([[[0.00065414]]], dtype=float32),\n", " array([[[0.00065419]]], dtype=float32),\n", " array([[[0.000654]]], dtype=float32),\n", " array([[[0.00065467]]], dtype=float32),\n", " array([[[0.00065543]]], dtype=float32),\n", " array([[[0.00065506]]], dtype=float32),\n", " array([[[0.00065486]]], dtype=float32),\n", " array([[[0.00065537]]], dtype=float32),\n", " array([[[0.00065632]]], dtype=float32),\n", " array([[[0.00065762]]], dtype=float32),\n", " array([[[0.0006574]]], dtype=float32),\n", " array([[[0.00065724]]], dtype=float32),\n", " array([[[0.00065767]]], dtype=float32),\n", " array([[[0.00065838]]], dtype=float32),\n", " array([[[0.00065982]]], dtype=float32),\n", " array([[[0.0006613]]], dtype=float32),\n", " array([[[0.00066154]]], dtype=float32),\n", " array([[[0.00066186]]], dtype=float32),\n", " array([[[0.00066264]]], dtype=float32),\n", " array([[[0.0006635]]], dtype=float32),\n", " array([[[0.00066396]]], dtype=float32),\n", " array([[[0.00066523]]], dtype=float32),\n", " array([[[0.00066677]]], dtype=float32),\n", " array([[[0.00066782]]], dtype=float32),\n", " array([[[0.00066901]]], dtype=float32),\n", " array([[[0.00066902]]], dtype=float32),\n", " array([[[0.00066898]]], dtype=float32),\n", " array([[[0.00066989]]], dtype=float32),\n", " array([[[0.00067137]]], dtype=float32),\n", " array([[[0.00067437]]], dtype=float32),\n", " array([[[0.00067679]]], dtype=float32),\n", " array([[[0.00067564]]], dtype=float32),\n", " array([[[0.00067558]]], dtype=float32),\n", " array([[[0.00067751]]], dtype=float32),\n", " array([[[0.00067914]]], dtype=float32),\n", " array([[[0.00067966]]], dtype=float32),\n", " array([[[0.00068042]]], dtype=float32),\n", " array([[[0.00068183]]], dtype=float32),\n", " array([[[0.00068292]]], dtype=float32),\n", " array([[[0.00068385]]], dtype=float32),\n", " array([[[0.00068445]]], dtype=float32),\n", " array([[[0.00068523]]], dtype=float32),\n", " array([[[0.00068666]]], dtype=float32),\n", " array([[[0.00068937]]], dtype=float32),\n", " array([[[0.00069001]]], dtype=float32),\n", " array([[[0.00068809]]], dtype=float32),\n", " array([[[0.00068873]]], dtype=float32),\n", " array([[[0.00069078]]], dtype=float32),\n", " array([[[0.00069214]]], dtype=float32),\n", " array([[[0.00069297]]], dtype=float32),\n", " array([[[0.0006936]]], dtype=float32),\n", " array([[[0.00069466]]], dtype=float32),\n", " array([[[0.00069676]]], dtype=float32),\n", " array([[[0.00069879]]], dtype=float32),\n", " array([[[0.00070022]]], dtype=float32),\n", " array([[[0.00070094]]], dtype=float32),\n", " array([[[0.00070159]]], dtype=float32),\n", " array([[[0.00070219]]], dtype=float32),\n", " array([[[0.00070397]]], dtype=float32),\n", " array([[[0.00070603]]], dtype=float32),\n", " array([[[0.00070573]]], dtype=float32),\n", " array([[[0.00070623]]], dtype=float32),\n", " array([[[0.00070761]]], dtype=float32),\n", " array([[[0.00070882]]], dtype=float32),\n", " array([[[0.00071004]]], dtype=float32),\n", " array([[[0.00071172]]], dtype=float32),\n", " array([[[0.00071414]]], dtype=float32),\n", " array([[[0.00071584]]], dtype=float32),\n", " array([[[0.00071549]]], dtype=float32),\n", " array([[[0.00071561]]], dtype=float32),\n", " array([[[0.00071647]]], dtype=float32),\n", " array([[[0.0007175]]], dtype=float32),\n", " array([[[0.00071925]]], dtype=float32),\n", " array([[[0.00072038]]], dtype=float32),\n", " array([[[0.00072055]]], dtype=float32),\n", " array([[[0.0007211]]], dtype=float32),\n", " array([[[0.00072235]]], dtype=float32),\n", " array([[[0.00072469]]], dtype=float32),\n", " array([[[0.00072703]]], dtype=float32),\n", " array([[[0.00072659]]], dtype=float32),\n", " array([[[0.00072634]]], dtype=float32),\n", " array([[[0.00072891]]], dtype=float32),\n", " array([[[0.0007307]]], dtype=float32),\n", " array([[[0.00073059]]], dtype=float32),\n", " array([[[0.00073057]]], dtype=float32),\n", " array([[[0.00073114]]], dtype=float32),\n", " array([[[0.00073237]]], dtype=float32),\n", " array([[[0.00073401]]], dtype=float32),\n", " array([[[0.00073588]]], dtype=float32),\n", " array([[[0.00073705]]], dtype=float32),\n", " array([[[0.00073682]]], dtype=float32),\n", " array([[[0.00073683]]], dtype=float32),\n", " array([[[0.00073886]]], dtype=float32),\n", " array([[[0.00074061]]], dtype=float32),\n", " array([[[0.00073967]]], dtype=float32),\n", " array([[[0.00073905]]], dtype=float32),\n", " array([[[0.00074018]]], dtype=float32),\n", " array([[[0.00074138]]], dtype=float32),\n", " array([[[0.0007417]]], dtype=float32),\n", " array([[[0.00074186]]], dtype=float32),\n", " array([[[0.00074224]]], dtype=float32),\n", " array([[[0.00074222]]], dtype=float32),\n", " array([[[0.0007423]]], dtype=float32),\n", " array([[[0.00074301]]], dtype=float32),\n", " array([[[0.00074369]]], dtype=float32),\n", " array([[[0.00074386]]], dtype=float32),\n", " array([[[0.00074339]]], dtype=float32),\n", " array([[[0.00074283]]], dtype=float32),\n", " array([[[0.00074315]]], dtype=float32),\n", " array([[[0.00074413]]], dtype=float32),\n", " array([[[0.0007446]]], dtype=float32),\n", " array([[[0.0007446]]], dtype=float32),\n", " array([[[0.00074443]]], dtype=float32),\n", " array([[[0.00074346]]], dtype=float32),\n", " array([[[0.00074249]]], dtype=float32),\n", " array([[[0.00074244]]], dtype=float32),\n", " array([[[0.0007431]]], dtype=float32),\n", " array([[[0.00074384]]], dtype=float32),\n", " array([[[0.0007438]]], dtype=float32),\n", " array([[[0.00074327]]], dtype=float32),\n", " array([[[0.00074336]]], dtype=float32),\n", " array([[[0.00074462]]], dtype=float32),\n", " array([[[0.00074635]]], dtype=float32),\n", " array([[[0.00074646]]], dtype=float32),\n", " array([[[0.00074524]]], dtype=float32),\n", " array([[[0.00074467]]], dtype=float32),\n", " array([[[0.00074474]]], dtype=float32),\n", " array([[[0.00074502]]], dtype=float32),\n", " array([[[0.00074483]]], dtype=float32),\n", " array([[[0.00074455]]], dtype=float32),\n", " array([[[0.00074484]]], dtype=float32),\n", " array([[[0.0007456]]], dtype=float32),\n", " array([[[0.00074677]]], dtype=float32),\n", " array([[[0.00074759]]], dtype=float32),\n", " array([[[0.00074722]]], dtype=float32),\n", " array([[[0.00074664]]], dtype=float32),\n", " array([[[0.00074644]]], dtype=float32),\n", " array([[[0.00074728]]], dtype=float32),\n", " array([[[0.00074906]]], dtype=float32),\n", " array([[[0.00074922]]], dtype=float32),\n", " array([[[0.00074799]]], dtype=float32),\n", " array([[[0.00074776]]], dtype=float32),\n", " array([[[0.00074889]]], dtype=float32),\n", " array([[[0.00074975]]], dtype=float32),\n", " array([[[0.00074996]]], dtype=float32),\n", " array([[[0.0007508]]], dtype=float32),\n", " array([[[0.00075228]]], dtype=float32),\n", " array([[[0.00075365]]], dtype=float32),\n", " array([[[0.00075319]]], dtype=float32),\n", " array([[[0.00075161]]], dtype=float32),\n", " array([[[0.00075122]]], dtype=float32),\n", " array([[[0.00075157]]], dtype=float32),\n", " array([[[0.00075232]]], dtype=float32),\n", " array([[[0.00075356]]], dtype=float32),\n", " array([[[0.00075435]]], dtype=float32),\n", " array([[[0.00075464]]], dtype=float32),\n", " array([[[0.00075505]]], dtype=float32),\n", " array([[[0.00075521]]], dtype=float32),\n", " array([[[0.00075534]]], dtype=float32),\n", " array([[[0.00075578]]], dtype=float32),\n", " array([[[0.00075674]]], dtype=float32),\n", " array([[[0.00075854]]], dtype=float32),\n", " array([[[0.00075967]]], dtype=float32),\n", " array([[[0.00075976]]], dtype=float32),\n", " array([[[0.00076007]]], dtype=float32),\n", " array([[[0.0007601]]], dtype=float32),\n", " array([[[0.00076023]]], dtype=float32),\n", " array([[[0.00076141]]], dtype=float32),\n", " array([[[0.00076229]]], dtype=float32),\n", " array([[[0.0007625]]], dtype=float32),\n", " array([[[0.00076309]]], dtype=float32),\n", " array([[[0.00076459]]], dtype=float32),\n", " array([[[0.000766]]], dtype=float32),\n", " array([[[0.0007659]]], dtype=float32),\n", " array([[[0.00076539]]], dtype=float32),\n", " array([[[0.00076602]]], dtype=float32),\n", " array([[[0.00076763]]], dtype=float32),\n", " array([[[0.00076944]]], dtype=float32),\n", " array([[[0.0007703]]], dtype=float32),\n", " array([[[0.00076999]]], dtype=float32),\n", " array([[[0.00077008]]], dtype=float32),\n", " array([[[0.00077127]]], dtype=float32),\n", " array([[[0.00077222]]], dtype=float32),\n", " array([[[0.00077225]]], dtype=float32),\n", " array([[[0.0007721]]], dtype=float32),\n", " array([[[0.00077221]]], dtype=float32),\n", " array([[[0.00077333]]], dtype=float32),\n", " array([[[0.00077663]]], dtype=float32),\n", " array([[[0.00078005]]], dtype=float32),\n", " array([[[0.0007797]]], dtype=float32),\n", " array([[[0.00077707]]], dtype=float32),\n", " array([[[0.00077615]]], dtype=float32),\n", " array([[[0.0007774]]], dtype=float32),\n", " array([[[0.0007792]]], dtype=float32),\n", " array([[[0.00078011]]], dtype=float32),\n", " array([[[0.00078073]]], dtype=float32),\n", " array([[[0.00078233]]], dtype=float32),\n", " array([[[0.00078476]]], dtype=float32),\n", " array([[[0.00078626]]], dtype=float32),\n", " array([[[0.00078597]]], dtype=float32),\n", " array([[[0.00078529]]], dtype=float32),\n", " array([[[0.00078508]]], dtype=float32),\n", " array([[[0.00078521]]], dtype=float32),\n", " array([[[0.00078609]]], dtype=float32),\n", " array([[[0.00078758]]], dtype=float32),\n", " array([[[0.00078923]]], dtype=float32),\n", " array([[[0.00079059]]], dtype=float32),\n", " array([[[0.00079102]]], dtype=float32),\n", " array([[[0.00079088]]], dtype=float32),\n", " array([[[0.00079099]]], dtype=float32),\n", " array([[[0.00079203]]], dtype=float32),\n", " array([[[0.00079415]]], dtype=float32),\n", " array([[[0.00079539]]], dtype=float32),\n", " array([[[0.00079461]]], dtype=float32),\n", " array([[[0.00079367]]], dtype=float32),\n", " array([[[0.0007938]]], dtype=float32),\n", " array([[[0.00079449]]], dtype=float32),\n", " array([[[0.00079561]]], dtype=float32),\n", " array([[[0.00079674]]], dtype=float32),\n", " array([[[0.00079697]]], dtype=float32),\n", " array([[[0.00079679]]], dtype=float32),\n", " array([[[0.00079728]]], dtype=float32),\n", " array([[[0.00079793]]], dtype=float32),\n", " array([[[0.00079771]]], dtype=float32),\n", " array([[[0.000797]]], dtype=float32),\n", " array([[[0.00079663]]], dtype=float32),\n", " array([[[0.00079674]]], dtype=float32),\n", " array([[[0.00079728]]], dtype=float32),\n", " array([[[0.00079791]]], dtype=float32),\n", " array([[[0.00079811]]], dtype=float32),\n", " array([[[0.00079774]]], dtype=float32),\n", " array([[[0.00079716]]], dtype=float32),\n", " array([[[0.00079686]]], dtype=float32),\n", " array([[[0.00079674]]], dtype=float32),\n", " array([[[0.00079648]]], dtype=float32),\n", " array([[[0.00079668]]], dtype=float32),\n", " array([[[0.00079828]]], dtype=float32),\n", " array([[[0.00080067]]], dtype=float32),\n", " array([[[0.00080206]]], dtype=float32),\n", " array([[[0.00080201]]], dtype=float32),\n", " array([[[0.00080103]]], dtype=float32),\n", " array([[[0.00079926]]], dtype=float32),\n", " array([[[0.00079787]]], dtype=float32),\n", " array([[[0.00079833]]], dtype=float32),\n", " array([[[0.00079969]]], dtype=float32),\n", " array([[[0.00080034]]], dtype=float32),\n", " array([[[0.00080043]]], dtype=float32),\n", " array([[[0.00080059]]], dtype=float32),\n", " array([[[0.00080095]]], dtype=float32),\n", " array([[[0.00080202]]], dtype=float32),\n", " array([[[0.00080435]]], dtype=float32),\n", " array([[[0.00080657]]], dtype=float32),\n", " array([[[0.00080686]]], dtype=float32),\n", " array([[[0.00080633]]], dtype=float32),\n", " array([[[0.00080695]]], dtype=float32),\n", " array([[[0.00080822]]], dtype=float32),\n", " array([[[0.00080934]]], dtype=float32),\n", " array([[[0.00081005]]], dtype=float32),\n", " array([[[0.00080953]]], dtype=float32),\n", " array([[[0.00080828]]], dtype=float32),\n", " array([[[0.00080809]]], dtype=float32),\n", " array([[[0.00080978]]], dtype=float32),\n", " array([[[0.00081238]]], dtype=float32),\n", " array([[[0.00081414]]], dtype=float32),\n", " array([[[0.00081489]]], dtype=float32),\n", " array([[[0.00081527]]], dtype=float32),\n", " array([[[0.00081508]]], dtype=float32),\n", " array([[[0.00081494]]], dtype=float32),\n", " array([[[0.00081574]]], dtype=float32),\n", " array([[[0.00081706]]], dtype=float32),\n", " array([[[0.00081882]]], dtype=float32),\n", " array([[[0.00082174]]], dtype=float32),\n", " array([[[0.00082502]]], dtype=float32),\n", " array([[[0.00082659]]], dtype=float32),\n", " array([[[0.00082614]]], dtype=float32),\n", " array([[[0.00082624]]], dtype=float32),\n", " array([[[0.00082881]]], dtype=float32),\n", " array([[[0.0008319]]], dtype=float32),\n", " array([[[0.00083251]]], dtype=float32),\n", " array([[[0.00083155]]], dtype=float32),\n", " array([[[0.0008321]]], dtype=float32),\n", " array([[[0.00083437]]], dtype=float32),\n", " array([[[0.00083639]]], dtype=float32),\n", " array([[[0.00083732]]], dtype=float32),\n", " array([[[0.00083755]]], dtype=float32),\n", " array([[[0.00083766]]], dtype=float32),\n", " array([[[0.00083851]]], dtype=float32),\n", " array([[[0.00084083]]], dtype=float32),\n", " array([[[0.000844]]], dtype=float32),\n", " array([[[0.00084633]]], dtype=float32),\n", " array([[[0.00084694]]], dtype=float32),\n", " array([[[0.00084688]]], dtype=float32),\n", " array([[[0.00084795]]], dtype=float32),\n", " array([[[0.00085118]]], dtype=float32),\n", " array([[[0.0008561]]], dtype=float32),\n", " array([[[0.00085976]]], dtype=float32),\n", " array([[[0.0008596]]], dtype=float32),\n", " array([[[0.00085722]]], dtype=float32),\n", " array([[[0.00085665]]], dtype=float32),\n", " array([[[0.00085927]]], dtype=float32),\n", " array([[[0.00086258]]], dtype=float32),\n", " array([[[0.00086357]]], dtype=float32),\n", " array([[[0.00086218]]], dtype=float32),\n", " array([[[0.00086124]]], dtype=float32),\n", " array([[[0.00086323]]], dtype=float32),\n", " array([[[0.00086827]]], dtype=float32),\n", " array([[[0.00087396]]], dtype=float32),\n", " array([[[0.00087677]]], dtype=float32),\n", " array([[[0.00087588]]], dtype=float32),\n", " array([[[0.00087414]]], dtype=float32),\n", " array([[[0.00087435]]], dtype=float32),\n", " array([[[0.00087654]]], dtype=float32),\n", " array([[[0.00088018]]], dtype=float32),\n", " array([[[0.00088456]]], dtype=float32),\n", " array([[[0.00088757]]], dtype=float32),\n", " array([[[0.00088784]]], dtype=float32),\n", " array([[[0.0008869]]], dtype=float32),\n", " array([[[0.00088683]]], dtype=float32),\n", " array([[[0.00088745]]], dtype=float32),\n", " array([[[0.00088795]]], dtype=float32),\n", " array([[[0.000889]]], dtype=float32),\n", " array([[[0.00089174]]], dtype=float32),\n", " array([[[0.00089624]]], dtype=float32),\n", " array([[[0.00090116]]], dtype=float32),\n", " array([[[0.00090419]]], dtype=float32),\n", " array([[[0.00090428]]], dtype=float32),\n", " array([[[0.00090319]]], dtype=float32),\n", " array([[[0.00090372]]], dtype=float32),\n", " array([[[0.00090636]]], dtype=float32),\n", " array([[[0.00090944]]], dtype=float32),\n", " array([[[0.00091174]]], dtype=float32),\n", " array([[[0.00091351]]], dtype=float32),\n", " array([[[0.00091572]]], dtype=float32),\n", " array([[[0.00091919]]], dtype=float32),\n", " array([[[0.00092398]]], dtype=float32),\n", " array([[[0.00092851]]], dtype=float32),\n", " array([[[0.00093059]]], dtype=float32),\n", " array([[[0.00092961]]], dtype=float32),\n", " array([[[0.00092764]]], dtype=float32),\n", " array([[[0.0009271]]], dtype=float32),\n", " array([[[0.00092823]]], dtype=float32),\n", " array([[[0.00092986]]], dtype=float32),\n", " array([[[0.00093136]]], dtype=float32),\n", " array([[[0.00093329]]], dtype=float32),\n", " array([[[0.00093595]]], dtype=float32),\n", " array([[[0.00093898]]], dtype=float32),\n", " array([[[0.00094228]]], dtype=float32),\n", " array([[[0.00094611]]], dtype=float32),\n", " array([[[0.00095031]]], dtype=float32),\n", " array([[[0.00095401]]], dtype=float32),\n", " array([[[0.00095662]]], dtype=float32),\n", " array([[[0.00095863]]], dtype=float32),\n", " array([[[0.00096089]]], dtype=float32),\n", " array([[[0.00096337]]], dtype=float32),\n", " array([[[0.0009652]]], dtype=float32),\n", " array([[[0.00096591]]], dtype=float32),\n", " array([[[0.00096619]]], dtype=float32),\n", " array([[[0.00096679]]], dtype=float32),\n", " array([[[0.0009675]]], dtype=float32),\n", " array([[[0.00096784]]], dtype=float32),\n", " array([[[0.00096783]]], dtype=float32),\n", " array([[[0.00096803]]], dtype=float32),\n", " array([[[0.0009687]]], dtype=float32),\n", " array([[[0.00096952]]], dtype=float32),\n", " array([[[0.00097034]]], dtype=float32),\n", " array([[[0.00097135]]], dtype=float32),\n", " array([[[0.00097248]]], dtype=float32),\n", " array([[[0.00097334]]], dtype=float32),\n", " array([[[0.00097377]]], dtype=float32),\n", " array([[[0.00097431]]], dtype=float32),\n", " array([[[0.00097539]]], dtype=float32),\n", " array([[[0.0009767]]], dtype=float32),\n", " array([[[0.00097769]]], dtype=float32),\n", " array([[[0.0009781]]], dtype=float32),\n", " array([[[0.00097804]]], dtype=float32),\n", " array([[[0.00097779]]], dtype=float32),\n", " array([[[0.0009779]]], dtype=float32),\n", " array([[[0.00097883]]], dtype=float32),\n", " array([[[0.00098018]]], dtype=float32),\n", " array([[[0.00098107]]], dtype=float32),\n", " array([[[0.00098114]]], dtype=float32),\n", " array([[[0.00098072]]], dtype=float32),\n", " array([[[0.00098019]]], dtype=float32),\n", " array([[[0.00097967]]], dtype=float32),\n", " array([[[0.00097928]]], dtype=float32),\n", " array([[[0.00097925]]], dtype=float32),\n", " array([[[0.00097985]]], dtype=float32),\n", " array([[[0.00098153]]], dtype=float32),\n", " array([[[0.00098495]]], dtype=float32),\n", " array([[[0.00099003]]], dtype=float32),\n", " array([[[0.00099501]]], dtype=float32),\n", " array([[[0.00099737]]], dtype=float32),\n", " array([[[0.00099587]]], dtype=float32),\n", " array([[[0.00099154]]], dtype=float32),\n", " array([[[0.00098665]]], dtype=float32),\n", " array([[[0.0009828]]], dtype=float32),\n", " array([[[0.0009806]]], dtype=float32),\n", " array([[[0.00098028]]], dtype=float32),\n", " array([[[0.00098177]]], dtype=float32),\n", " array([[[0.00098453]]], dtype=float32),\n", " array([[[0.00098773]]], dtype=float32),\n", " array([[[0.00099064]]], dtype=float32),\n", " array([[[0.00099284]]], dtype=float32),\n", " array([[[0.00099426]]], dtype=float32),\n", " array([[[0.0009952]]], dtype=float32),\n", " array([[[0.00099584]]], dtype=float32),\n", " array([[[0.00099601]]], dtype=float32),\n", " array([[[0.00099583]]], dtype=float32),\n", " array([[[0.0009961]]], dtype=float32),\n", " array([[[0.00099773]]], dtype=float32),\n", " array([[[0.00100047]]], dtype=float32),\n", " array([[[0.00100289]]], dtype=float32),\n", " array([[[0.00100365]]], dtype=float32),\n", " array([[[0.00100267]]], dtype=float32),\n", " array([[[0.0010008]]], dtype=float32),\n", " array([[[0.00099888]]], dtype=float32),\n", " array([[[0.00099734]]], dtype=float32),\n", " array([[[0.00099641]]], dtype=float32),\n", " array([[[0.00099633]]], dtype=float32),\n", " array([[[0.00099737]]], dtype=float32),\n", " array([[[0.00099961]]], dtype=float32),\n", " array([[[0.00100282]]], dtype=float32),\n", " array([[[0.00100632]]], dtype=float32),\n", " array([[[0.00100895]]], dtype=float32),\n", " array([[[0.00100968]]], dtype=float32),\n", " array([[[0.0010083]]], dtype=float32),\n", " array([[[0.00100585]]], dtype=float32),\n", " array([[[0.00100405]]], dtype=float32),\n", " array([[[0.0010042]]], dtype=float32),\n", " array([[[0.00100633]]], dtype=float32),\n", " array([[[0.00100958]]], dtype=float32),\n", " array([[[0.00101298]]], dtype=float32),\n", " array([[[0.00101566]]], dtype=float32),\n", " array([[[0.00101689]]], dtype=float32),\n", " array([[[0.00101661]]], dtype=float32),\n", " array([[[0.00101568]]], dtype=float32),\n", " array([[[0.00101546]]], dtype=float32),\n", " array([[[0.00101705]]], dtype=float32),\n", " array([[[0.00102068]]], dtype=float32),\n", " array([[[0.00102545]]], dtype=float32),\n", " array([[[0.00102955]]], dtype=float32),\n", " array([[[0.00103113]]], dtype=float32),\n", " array([[[0.00102966]]], dtype=float32),\n", " array([[[0.00102646]]], dtype=float32),\n", " array([[[0.00102365]]], dtype=float32),\n", " ...]}" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "lightcurve_1d.results" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "F_jpbrtgQIdq" }, "source": [ "but this output is quite ugly. Instead, you can call some routines to make informative summaries of them that are easier to read" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "UaaEuY1nQNv_", "outputId": "76ff2c82-b246-454c-940f-5c17568c9770" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "loss: [1.089931]\n", "delta_loss: [5.364418e-06]\n", "likelihood.second_noise_covar.noise: [0.42528677]\n", "mean_module.constant: [-0.03767705]\n", "covar_module.mixture_weights: [0.00633732]\n", "covar_module.mixture_means: [0.07724167]\n", "covar_module.mixture_scales: [0.00486984]\n" ] } ], "source": [ "lightcurve_1d.print_results()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Period 0: [12.946379] weight: 0.006337318103760481\n" ] } ], "source": [ "lightcurve_1d.print_periods()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "vmeUoeWnQn5_" }, "source": [ "and you can also generate some plots of the output; for example the lightcurve with the fit results on top of it" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 546 }, "id": "SSqM2r0RQvnz", "outputId": "dd5cda6e-ebdb-4b95-ea19-23a3592ca641" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig = lightcurve_1d.plot(save=False)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": { "id": "ZQHFjRrAQzsD" }, "source": [ "or the maximum a-posteriori PSD that pgmuvi has inferred (this can be very slow!)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 169 }, "id": "CzIGiIXKQ5Td", "outputId": "4271fa98-6e62-4c1c-d48a-db81dc03572a" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "tensor([0.0010], grad_fn=) 0.25 8.0\n", "torch.Size([8214])\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "lightcurve_1d.plot_psd()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "However, this fit wasn't particularly good! This is because we didn't specify any initial guesses for the hyperparameters, and the default values are not very good for this dataset. We can try again with some better guesses:" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "likelihood.second_noise_covar.noise: tensor([0.4253])\n", "mean_module.constant: -0.03767704963684082\n", "covar_module.mixture_weights: tensor([0.0063])\n", "covar_module.mixture_means: tensor([[[0.0772]]])\n", "covar_module.mixture_scales: tensor([[[0.0049]]])\n", "196.48998803679268\n", "likelihood.second_noise_covar.noise: tensor([0.1000])\n", "mean_module.constant: -0.03767704963684082\n", "covar_module.mixture_weights: tensor([0.0063])\n", "covar_module.mixture_means: tensor([[[0.0051]]])\n", "covar_module.mixture_scales: tensor([[[0.0049]]])\n" ] } ], "source": [ "lightcurve_1d.print_parameters()\n", "print(period_guess)\n", "guess = {\n", " \"likelihood.second_noise_covar.noise\": torch.Tensor([0.1]),\n", " \"sci_kernel.mixture_means\": torch.Tensor([1 / period_guess]),\n", "}\n", "lightcurve_1d.set_hypers(guess)\n", "lightcurve_1d.print_parameters()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We printed the parameters before and after setting them to see how they have changed. As we can see, they were updated as instructed. Now we can try to fit again:" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "likelihood.second_noise_covar.noise: tensor([0.1000])\n", "mean_module.constant: -0.03767704963684082\n", "covar_module.mixture_weights: tensor([0.0063])\n", "covar_module.mixture_means: tensor([[[0.0051]]])\n", "covar_module.mixture_scales: tensor([[[0.0049]]])\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 3000/3000 [00:40<00:00, 74.22it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "loss: [-0.36470833]\n", "delta_loss: [1.6003847e-05]\n", "likelihood.second_noise_covar.noise: [0.00239522]\n", "mean_module.constant: [0.00569272]\n", "covar_module.mixture_weights: [0.47454086]\n", "covar_module.mixture_means: [0.00560915]\n", "covar_module.mixture_scales: [1.4601766e-06]\n", "Period 0: [178.2802] weight: 0.4745408594608307\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "new_fit = lightcurve_1d.fit(training_iter=3000, miniter=3000)\n", "lightcurve_1d.print_results()\n", "lightcurve_1d.print_periods()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "This results in a much better fit, as you can see:" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig = lightcurve_1d.plot(save=False)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We can look at the new PSD, but again this can take a very long time!" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "tensor([2.9204e-07], grad_fn=) 0.25 8.0\n", "torch.Size([27391534])\n", "torch.Size([27391534]) torch.Size([1, 1, 1]) torch.Size([1, 1, 1]) torch.Size([1])\n", "Normal(loc: tensor([[[0.0056]]]), scale: tensor([[[1.4602e-06]]]))\n", "torch.Size([27391534])\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "lightcurve_1d.plot_psd(log=(True, True), truncate_psd=True, debug=True)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We can also look at the periods that came out of it:" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Periods: tensor([[178.2802]]), Weights: tensor([0.4745]), Scales: tensor([[108997.0547]])\n" ] } ], "source": [ "periods, weights, scales = lightcurve_1d.get_periods()\n", "print(f\"Periods: {periods}, Weights: {weights}, Scales: {scales}\")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "And look at the frequency that we used as our guess of the period:" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.0050893178323811105" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" } ], "source": [ "1 / period_guess" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Now we will take a look at another slightly more advanced topic - setting constraints on the parameters. \n", "`pgmuvi` tries to make this easy for you, by providing a convenient method `lightcurve1d.set_constraint()`. \n", "However, there are a few caveats in using it.\n", "\n", "Adjusting the constraints on the parameters will not change the values of the internal representation that `gpytorch` uses for the parameters. \n", "*This will lead to the seemingly odd behaviour that updating the constraints will change the reported values of the parameters!*\n", "This is because `gpytorch` works with an internal representation of the parameters in an unconstrained space, and the constraints are really a transformation from the unconstrained space to a constrained space.\n", "This is done because the unconstrained space is easier to optimise in, and the constraints are applied to transform them to values that are meaningful for the model.\n", "As a result, if you want to update the constraints, **you must set the values of the parameters again afterwards!**\n", "Perhaps even more importantly, if you are updating constraints after performing a fit (as we are here) and want to keep the final state of the parameters, you should save the final state of the parameters, update the constraints, and then set the parameters to the saved values. \n", "To demonstrate this, we will go through a few of the examples of this here." ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'likelihood.second_noise_covar.noise': tensor([0.0024]), 'mean_module.constant': tensor(0.0057), 'covar_module.mixture_weights': tensor([0.4745]), 'covar_module.mixture_means': tensor([[[0.0056]]]), 'covar_module.mixture_scales': tensor([[[1.4602e-06]]])}\n" ] } ], "source": [ "# first, let's save the final values of the parameters\n", "pars = lightcurve_1d.get_parameters()\n", "print(pars)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Now we can set some constraints on the parameters. \n", "If you're running this yourself, remember to check that the ranges of the constraints include the values of the parameters that were found in the fit!\n", "Otherwise you will get an error when you try to set the parameters to the values found in the fit." ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Found parameter likelihood.second_noise_covar.noise in model parameters\n", "Parameter likelihood.second_noise_covar.noise will have constraint: Interval(1.000E-03, 1.500E-01)\n", "which may be transformed\n", "Interval(1.000E-03, 1.500E-01)\n", "Registered constraint Interval(1.000E-03, 1.500E-01)\n", "to parameter likelihood.second_noise_covar.noise\n", "Found parameter mean_module.constant in model parameters\n", "Parameter mean_module.constant will have constraint: Interval(-1.000E-01, 1.000E-01)\n", "which may be transformed\n", "Interval(-1.000E-01, 1.000E-01)\n", "Registered constraint Interval(-1.000E-01, 1.000E-01)\n", "to parameter mean_module.constant\n", "Found parameter covar_module.mixture_means in model parameters\n", "Parameter covar_module.mixture_means will have constraint: Interval(1.000E-03, 1.000E-01)\n", "which may be transformed\n", "Interval(1.000E-03, 1.000E-01)\n", "tensor([1.4183])\n", "tensor(1.4183)\n", "Interval(1.418E+00, 1.000E-01)\n", "tensor(141.8250)\n", "Interval(1.418E+00, 1.418E+02)\n", "Interval(1.418E+00, 1.418E+02)\n", "Registered constraint Interval(1.418E+00, 1.418E+02)\n", "to parameter covar_module.mixture_means\n", "Found parameter covar_module.mixture_scales in model parameters\n", "Parameter covar_module.mixture_scales will have constraint: Interval(1.000E-06, 2.000E-01)\n", "which may be transformed\n", "Interval(1.000E-06, 2.000E-01)\n", "tensor([0.0014])\n", "tensor(0.0014)\n", "Interval(1.418E-03, 2.000E-01)\n", "tensor(283.6500)\n", "Interval(1.418E-03, 2.836E+02)\n", "Interval(1.418E-03, 2.836E+02)\n", "Registered constraint Interval(1.418E-03, 2.836E+02)\n", "to parameter covar_module.mixture_scales\n" ] } ], "source": [ "# now, let's update some constraints!\n", "const_dict = {\n", " \"likelihood.second_noise_covar.noise\": gpytorch.constraints.Interval(0.001, 0.15),\n", " \"mean_module.constant\": gpytorch.constraints.Interval(-0.1, 0.1),\n", " \"covar_module.mixture_means\": gpytorch.constraints.Interval(1 / 1000.0, 1 / 10.0),\n", " \"covar_module.mixture_scales\": gpytorch.constraints.Interval(1e-6, 0.2),\n", "}\n", "lightcurve_1d.set_constraint(const_dict, debug=True)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Now we've updated the constraints, let's see what has happened to the parameters:" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "likelihood.second_noise_covar.noise: tensor([0.0015])\n", "mean_module.constant: -0.00119888037443161\n", "covar_module.mixture_weights: tensor([0.4745])\n", "covar_module.mixture_means: tensor([[[0.0999]]])\n", "covar_module.mixture_scales: tensor([[[0.0004]]])\n" ] } ], "source": [ "lightcurve_1d.print_parameters()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We can see these are different to those we found above. Let's see what happens if we set the parameters to the values we found above:" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "hypers before transform:\n", "{'likelihood.second_noise_covar.noise': tensor([0.0024]), 'mean_module.constant': tensor(0.0057), 'covar_module.mixture_weights': tensor([0.4745]), 'covar_module.mixture_means': tensor([[[0.0056]]]), 'covar_module.mixture_scales': tensor([[[1.4602e-06]]])}\n", "Applying x-transform to covar_module.mixture_means\n", "Applying x-transform to covar_module.mixture_scales\n", "hypers after transform:\n", "{'likelihood.second_noise_covar.noise': tensor([0.0024]), 'mean_module.constant': tensor(0.0057), 'covar_module.mixture_weights': tensor([0.4745]), 'covar_module.mixture_means': tensor([[[7.9552]]]), 'covar_module.mixture_scales': tensor([[[0.0021]]])}\n", "likelihood.second_noise_covar.noise: tensor([0.0024])\n", "mean_module.constant: 0.005692720413208008\n", "covar_module.mixture_weights: tensor([0.4745])\n", "covar_module.mixture_means: tensor([[[0.0056]]])\n", "covar_module.mixture_scales: tensor([[[1.4602e-06]]])\n" ] } ], "source": [ "lightcurve_1d.set_hypers(pars, debug=True)\n", "lightcurve_1d.print_parameters()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Now they should be back to normal! Let's see what happens if we try to fit again:" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "likelihood.second_noise_covar.noise: tensor([0.0069])\n", "mean_module.constant: 0.08803987503051758\n", "covar_module.mixture_weights: tensor([0.4745])\n", "covar_module.mixture_means: tensor([[[0.0007]]])\n", "covar_module.mixture_scales: tensor([[[1.4602e-06]]])\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 3000/3000 [00:40<00:00, 73.71it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "loss: [-0.3644622]\n", "delta_loss: [-2.2947788e-06]\n", "likelihood.second_noise_covar.noise: [0.00245486]\n", "mean_module.constant: [0.00537109]\n", "covar_module.mixture_weights: [0.43169224]\n", "covar_module.mixture_means: [0.00560879]\n", "covar_module.mixture_scales: [3.051211e-06]\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "fit = lightcurve_1d.fit(training_iter=3000, miniter=3000)\n", "lightcurve_1d.print_results()" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig = lightcurve_1d.plot()" ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(tensor([[178.2915]]), tensor([0.4317]), tensor([[52161.2383]]))" ] }, "execution_count": 23, "metadata": {}, "output_type": "execute_result" } ], "source": [ "lightcurve_1d.get_periods()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "You can see that the best-fit period has changed slightly. This is because the constraints we set are not exactly the same as the ones we used in the fit, so the optimiser is not able to find exactly the same solution. However, it is very close, and the fit is still very good." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "This is the end of this example. You can find more information in the other example notebooks, for example on how to use MCMC to sample the posterior distribution of the parameters, or how to fit multi-wavelength lightcurves." ] } ], "metadata": { "colab": { "include_colab_link": true, "provenance": [] }, "kernelspec": { "display_name": "Python 3", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.4" } }, "nbformat": 4, "nbformat_minor": 0 }