Datasets
Models
Downloads: 1
Switch to unified view
| a | b/Code/All Qiskit, PennyLane QML Nov 23/09a Qiskit ML Model 90%Test kkawchak.ipynb | ||
|---|---|---|---|
| 1 | { |
||
| 2 | "cells": [ |
||
| 3 | { |
||
| 4 | "cell_type": "markdown", |
||
| 5 | "id": "measured-liabilities", |
||
| 6 | "metadata": {}, |
||
| 7 | "source": [ |
||
| 8 | "# Saving, Loading Qiskit Machine Learning Models and Continuous Training\n", |
||
| 9 | "\n", |
||
| 10 | "In this tutorial we will show how to save and load Qiskit machine learning models. Ability to save a model is very important, especially when a significant amount of time is invested in training a model on a real hardware. Also, we will show how to resume training of the previously saved model.\n", |
||
| 11 | "\n", |
||
| 12 | "In this tutorial we will cover how to:\n", |
||
| 13 | "\n", |
||
| 14 | "* Generate a simple dataset, split it into training/test datasets and plot them\n", |
||
| 15 | "* Train and save a model\n", |
||
| 16 | "* Load a saved model and resume training\n", |
||
| 17 | "* Evaluate performance of models\n", |
||
| 18 | "* PyTorch hybrid models" |
||
| 19 | ] |
||
| 20 | }, |
||
| 21 | { |
||
| 22 | "cell_type": "markdown", |
||
| 23 | "id": "speaking-glance", |
||
| 24 | "metadata": {}, |
||
| 25 | "source": [ |
||
| 26 | "First off, we start from the required imports. We'll heavily use SciKit-Learn on the data preparation step. In the next cell we also fix a random seed for reproducibility purposes." |
||
| 27 | ] |
||
| 28 | }, |
||
| 29 | { |
||
| 30 | "cell_type": "code", |
||
| 31 | "execution_count": 236, |
||
| 32 | "id": "exposed-cholesterol", |
||
| 33 | "metadata": {}, |
||
| 34 | "outputs": [], |
||
| 35 | "source": [ |
||
| 36 | "import matplotlib.pyplot as plt\n", |
||
| 37 | "import numpy as np\n", |
||
| 38 | "from qiskit.algorithms.optimizers import COBYLA, ADAM, AQGD\n", |
||
| 39 | "from qiskit.circuit.library import RealAmplitudes\n", |
||
| 40 | "from qiskit.primitives import Sampler\n", |
||
| 41 | "from qiskit.utils import algorithm_globals\n", |
||
| 42 | "from sklearn.model_selection import train_test_split\n", |
||
| 43 | "from sklearn.preprocessing import OneHotEncoder, MinMaxScaler\n", |
||
| 44 | "\n", |
||
| 45 | "from qiskit_machine_learning.algorithms.classifiers import VQC\n", |
||
| 46 | "\n", |
||
| 47 | "from IPython.display import clear_output\n", |
||
| 48 | "\n", |
||
| 49 | "algorithm_globals.random_seed = 42" |
||
| 50 | ] |
||
| 51 | }, |
||
| 52 | { |
||
| 53 | "cell_type": "markdown", |
||
| 54 | "id": "rural-mileage", |
||
| 55 | "metadata": {}, |
||
| 56 | "source": [ |
||
| 57 | "We will be using two quantum simulators, in particular, two instances of the `Sampler` primitive. We'll start training on the first one, then will resume training on the second one. The approach shown in this tutorial can be used to train a model on a real hardware available on the cloud and then re-use the model for inference on a local simulator." |
||
| 58 | ] |
||
| 59 | }, |
||
| 60 | { |
||
| 61 | "cell_type": "code", |
||
| 62 | "execution_count": 237, |
||
| 63 | "id": "charming-seating", |
||
| 64 | "metadata": {}, |
||
| 65 | "outputs": [], |
||
| 66 | "source": [ |
||
| 67 | "sampler1 = Sampler()\n", |
||
| 68 | "\n", |
||
| 69 | "sampler2 = Sampler()" |
||
| 70 | ] |
||
| 71 | }, |
||
| 72 | { |
||
| 73 | "cell_type": "markdown", |
||
| 74 | "id": "careful-allowance", |
||
| 75 | "metadata": {}, |
||
| 76 | "source": [ |
||
| 77 | "## 1. Prepare a dataset\n", |
||
| 78 | "\n", |
||
| 79 | "Next step is to prepare a dataset. Here, we generate some data in the same way as in other tutorials. The difference is that we apply some transformations to the generated data. We generates `40` samples, each sample has `2` features, so our features is an array of shape `(40, 2)`. Labels are obtained by summing up features by columns and if the sum is more than `1` then this sample is labeled as `1` and `0` otherwise." |
||
| 80 | ] |
||
| 81 | }, |
||
| 82 | { |
||
| 83 | "cell_type": "code", |
||
| 84 | "execution_count": 238, |
||
| 85 | "id": "ceramic-florida", |
||
| 86 | "metadata": {}, |
||
| 87 | "outputs": [], |
||
| 88 | "source": [ |
||
| 89 | "num_samples = 80\n", |
||
| 90 | "num_features = 2\n", |
||
| 91 | "features = 2 * algorithm_globals.random.random([num_samples, num_features]) - 1\n", |
||
| 92 | "labels = 1 * (np.sum(features, axis=1) >= 0) # in { 0, 1}" |
||
| 93 | ] |
||
| 94 | }, |
||
| 95 | { |
||
| 96 | "cell_type": "markdown", |
||
| 97 | "id": "reduced-injury", |
||
| 98 | "metadata": {}, |
||
| 99 | "source": [ |
||
| 100 | "Then, we scale down our features into a range of `[0, 1]` by applying `MinMaxScaler` from SciKit-Learn. Model training convergence is better when this transformation is applied." |
||
| 101 | ] |
||
| 102 | }, |
||
| 103 | { |
||
| 104 | "cell_type": "code", |
||
| 105 | "execution_count": 239, |
||
| 106 | "id": "dirty-director", |
||
| 107 | "metadata": {}, |
||
| 108 | "outputs": [ |
||
| 109 | { |
||
| 110 | "data": { |
||
| 111 | "text/plain": [ |
||
| 112 | "(80, 2)" |
||
| 113 | ] |
||
| 114 | }, |
||
| 115 | "execution_count": 239, |
||
| 116 | "metadata": {}, |
||
| 117 | "output_type": "execute_result" |
||
| 118 | } |
||
| 119 | ], |
||
| 120 | "source": [ |
||
| 121 | "features = MinMaxScaler().fit_transform(features)\n", |
||
| 122 | "features.shape" |
||
| 123 | ] |
||
| 124 | }, |
||
| 125 | { |
||
| 126 | "cell_type": "markdown", |
||
| 127 | "id": "julian-amount", |
||
| 128 | "metadata": {}, |
||
| 129 | "source": [ |
||
| 130 | "Let's take a look at the features of the first `5` samples of our dataset after the transformation." |
||
| 131 | ] |
||
| 132 | }, |
||
| 133 | { |
||
| 134 | "cell_type": "code", |
||
| 135 | "execution_count": 240, |
||
| 136 | "id": "thorough-script", |
||
| 137 | "metadata": {}, |
||
| 138 | "outputs": [ |
||
| 139 | { |
||
| 140 | "data": { |
||
| 141 | "text/plain": [ |
||
| 142 | "array([[0.79246319, 0.44566143],\n", |
||
| 143 | " [0.88174919, 0.7126244 ],\n", |
||
| 144 | " [0.07538643, 1. ],\n", |
||
| 145 | " [0.77894364, 0.80422817],\n", |
||
| 146 | " [0.11118473, 0.45754615]])" |
||
| 147 | ] |
||
| 148 | }, |
||
| 149 | "execution_count": 240, |
||
| 150 | "metadata": {}, |
||
| 151 | "output_type": "execute_result" |
||
| 152 | } |
||
| 153 | ], |
||
| 154 | "source": [ |
||
| 155 | "features[0:5, :]" |
||
| 156 | ] |
||
| 157 | }, |
||
| 158 | { |
||
| 159 | "cell_type": "markdown", |
||
| 160 | "id": "racial-aluminum", |
||
| 161 | "metadata": {}, |
||
| 162 | "source": [ |
||
| 163 | "We choose `VQC` or Variational Quantum Classifier as a model we will train. This model, by default, takes one-hot encoded labels, so we have to transform the labels that are in the set of `{0, 1}` into one-hot representation. We employ SciKit-Learn for this transformation as well. Please note that the input array must be reshaped to `(num_samples, 1)` first. The `OneHotEncoder` encoder does not work with 1D arrays and our labels is a 1D array. In this case a user must decide either an array has only one feature(our case!) or has one sample. Also, by default the encoder returns sparse arrays, but for dataset plotting it is easier to have dense arrays, so we set `sparse` to `False`. " |
||
| 164 | ] |
||
| 165 | }, |
||
| 166 | { |
||
| 167 | "cell_type": "code", |
||
| 168 | "execution_count": 241, |
||
| 169 | "id": "understood-ukraine", |
||
| 170 | "metadata": {}, |
||
| 171 | "outputs": [ |
||
| 172 | { |
||
| 173 | "name": "stderr", |
||
| 174 | "output_type": "stream", |
||
| 175 | "text": [ |
||
| 176 | "/opt/conda/lib/python3.10/site-packages/sklearn/preprocessing/_encoders.py:828: FutureWarning: `sparse` was renamed to `sparse_output` in version 1.2 and will be removed in 1.4. `sparse_output` is ignored unless you leave `sparse` to its default value.\n", |
||
| 177 | " warnings.warn(\n" |
||
| 178 | ] |
||
| 179 | }, |
||
| 180 | { |
||
| 181 | "data": { |
||
| 182 | "text/plain": [ |
||
| 183 | "(80, 2)" |
||
| 184 | ] |
||
| 185 | }, |
||
| 186 | "execution_count": 241, |
||
| 187 | "metadata": {}, |
||
| 188 | "output_type": "execute_result" |
||
| 189 | } |
||
| 190 | ], |
||
| 191 | "source": [ |
||
| 192 | "labels = OneHotEncoder(sparse=False).fit_transform(labels.reshape(-1, 1))\n", |
||
| 193 | "labels.shape" |
||
| 194 | ] |
||
| 195 | }, |
||
| 196 | { |
||
| 197 | "cell_type": "markdown", |
||
| 198 | "id": "statewide-symbol", |
||
| 199 | "metadata": {}, |
||
| 200 | "source": [ |
||
| 201 | "Let's take a look at the labels of the first `5` labels of the dataset. The labels should be one-hot encoded." |
||
| 202 | ] |
||
| 203 | }, |
||
| 204 | { |
||
| 205 | "cell_type": "code", |
||
| 206 | "execution_count": 242, |
||
| 207 | "id": "german-agreement", |
||
| 208 | "metadata": {}, |
||
| 209 | "outputs": [ |
||
| 210 | { |
||
| 211 | "data": { |
||
| 212 | "text/plain": [ |
||
| 213 | "array([[0., 1.],\n", |
||
| 214 | " [0., 1.],\n", |
||
| 215 | " [0., 1.],\n", |
||
| 216 | " [0., 1.],\n", |
||
| 217 | " [1., 0.]])" |
||
| 218 | ] |
||
| 219 | }, |
||
| 220 | "execution_count": 242, |
||
| 221 | "metadata": {}, |
||
| 222 | "output_type": "execute_result" |
||
| 223 | } |
||
| 224 | ], |
||
| 225 | "source": [ |
||
| 226 | "labels[0:5, :]" |
||
| 227 | ] |
||
| 228 | }, |
||
| 229 | { |
||
| 230 | "cell_type": "markdown", |
||
| 231 | "id": "aquatic-toner", |
||
| 232 | "metadata": {}, |
||
| 233 | "source": [ |
||
| 234 | "Now we split our dataset into two parts: a training dataset and a test one. As a rule of thumb, 80% of a full dataset should go into a training part and 20% into a test one. Our training dataset has `30` samples. The test dataset should be used only once, when a model is trained to verify how well the model behaves on unseen data. We employ `train_test_split` from SciKit-Learn." |
||
| 235 | ] |
||
| 236 | }, |
||
| 237 | { |
||
| 238 | "cell_type": "code", |
||
| 239 | "execution_count": 243, |
||
| 240 | "id": "about-ordinary", |
||
| 241 | "metadata": {}, |
||
| 242 | "outputs": [ |
||
| 243 | { |
||
| 244 | "data": { |
||
| 245 | "text/plain": [ |
||
| 246 | "(60, 2)" |
||
| 247 | ] |
||
| 248 | }, |
||
| 249 | "execution_count": 243, |
||
| 250 | "metadata": {}, |
||
| 251 | "output_type": "execute_result" |
||
| 252 | } |
||
| 253 | ], |
||
| 254 | "source": [ |
||
| 255 | "train_features, test_features, train_labels, test_labels = train_test_split(\n", |
||
| 256 | " features, labels, train_size=60, random_state=algorithm_globals.random_seed\n", |
||
| 257 | ")\n", |
||
| 258 | "train_features.shape" |
||
| 259 | ] |
||
| 260 | }, |
||
| 261 | { |
||
| 262 | "cell_type": "markdown", |
||
| 263 | "id": "critical-angel", |
||
| 264 | "metadata": {}, |
||
| 265 | "source": [ |
||
| 266 | "Now it is time to see how our dataset looks like. Let's plot it." |
||
| 267 | ] |
||
| 268 | }, |
||
| 269 | { |
||
| 270 | "cell_type": "code", |
||
| 271 | "execution_count": 244, |
||
| 272 | "id": "fifty-scottish", |
||
| 273 | "metadata": {}, |
||
| 274 | "outputs": [ |
||
| 275 | { |
||
| 276 | "data": { |
||
| 277 | "image/png": "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\n", |
||
| 278 | "text/plain": [ |
||
| 279 | "<Figure size 1200x600 with 1 Axes>" |
||
| 280 | ] |
||
| 281 | }, |
||
| 282 | "metadata": {}, |
||
| 283 | "output_type": "display_data" |
||
| 284 | } |
||
| 285 | ], |
||
| 286 | "source": [ |
||
| 287 | "def plot_dataset():\n", |
||
| 288 | " plt.scatter(\n", |
||
| 289 | " train_features[np.where(train_labels[:, 0] == 0), 0],\n", |
||
| 290 | " train_features[np.where(train_labels[:, 0] == 0), 1],\n", |
||
| 291 | " marker=\"o\",\n", |
||
| 292 | " color=\"b\",\n", |
||
| 293 | " label=\"Label 0 train\",\n", |
||
| 294 | " )\n", |
||
| 295 | " plt.scatter(\n", |
||
| 296 | " train_features[np.where(train_labels[:, 0] == 1), 0],\n", |
||
| 297 | " train_features[np.where(train_labels[:, 0] == 1), 1],\n", |
||
| 298 | " marker=\"o\",\n", |
||
| 299 | " color=\"g\",\n", |
||
| 300 | " label=\"Label 1 train\",\n", |
||
| 301 | " )\n", |
||
| 302 | "\n", |
||
| 303 | " plt.scatter(\n", |
||
| 304 | " test_features[np.where(test_labels[:, 0] == 0), 0],\n", |
||
| 305 | " test_features[np.where(test_labels[:, 0] == 0), 1],\n", |
||
| 306 | " marker=\"o\",\n", |
||
| 307 | " facecolors=\"w\",\n", |
||
| 308 | " edgecolors=\"b\",\n", |
||
| 309 | " label=\"Label 0 test\",\n", |
||
| 310 | " )\n", |
||
| 311 | " plt.scatter(\n", |
||
| 312 | " test_features[np.where(test_labels[:, 0] == 1), 0],\n", |
||
| 313 | " test_features[np.where(test_labels[:, 0] == 1), 1],\n", |
||
| 314 | " marker=\"o\",\n", |
||
| 315 | " facecolors=\"w\",\n", |
||
| 316 | " edgecolors=\"g\",\n", |
||
| 317 | " label=\"Label 1 test\",\n", |
||
| 318 | " )\n", |
||
| 319 | "\n", |
||
| 320 | " plt.legend(bbox_to_anchor=(1.05, 1), loc=\"upper left\", borderaxespad=0.0)\n", |
||
| 321 | " plt.plot([1, 0], [0, 1], \"--\", color=\"black\")\n", |
||
| 322 | "\n", |
||
| 323 | "\n", |
||
| 324 | "plot_dataset()\n", |
||
| 325 | "plt.show()" |
||
| 326 | ] |
||
| 327 | }, |
||
| 328 | { |
||
| 329 | "cell_type": "markdown", |
||
| 330 | "id": "regulation-depression", |
||
| 331 | "metadata": {}, |
||
| 332 | "source": [ |
||
| 333 | "On the plot above we see:\n", |
||
| 334 | "\n", |
||
| 335 | "* Solid <span style=\"color:blue\">blue</span> dots are the samples from the training dataset labeled as `0`\n", |
||
| 336 | "* Empty <span style=\"color:blue\">blue</span> dots are the samples from the test dataset labeled as `0`\n", |
||
| 337 | "* Solid <span style=\"color:green\">green</span> dots are the samples from the training dataset labeled as `1`\n", |
||
| 338 | "* Empty <span style=\"color:green\">green</span> dots are the samples from the test dataset labeled as `1`\n", |
||
| 339 | "\n", |
||
| 340 | "We'll train our model using solid dots and verify it using empty dots." |
||
| 341 | ] |
||
| 342 | }, |
||
| 343 | { |
||
| 344 | "cell_type": "markdown", |
||
| 345 | "id": "egyptian-campaign", |
||
| 346 | "metadata": {}, |
||
| 347 | "source": [ |
||
| 348 | "## 2. Train a model and save it\n", |
||
| 349 | "\n", |
||
| 350 | "We'll train our model in two steps. On the first step we train our model in `20` iterations." |
||
| 351 | ] |
||
| 352 | }, |
||
| 353 | { |
||
| 354 | "cell_type": "code", |
||
| 355 | "execution_count": 245, |
||
| 356 | "id": "brief-lending", |
||
| 357 | "metadata": {}, |
||
| 358 | "outputs": [], |
||
| 359 | "source": [ |
||
| 360 | "maxiter = 20" |
||
| 361 | ] |
||
| 362 | }, |
||
| 363 | { |
||
| 364 | "cell_type": "markdown", |
||
| 365 | "id": "crude-franklin", |
||
| 366 | "metadata": {}, |
||
| 367 | "source": [ |
||
| 368 | "Create an empty array for callback to store values of the objective function." |
||
| 369 | ] |
||
| 370 | }, |
||
| 371 | { |
||
| 372 | "cell_type": "code", |
||
| 373 | "execution_count": 246, |
||
| 374 | "id": "integrated-palestinian", |
||
| 375 | "metadata": {}, |
||
| 376 | "outputs": [], |
||
| 377 | "source": [ |
||
| 378 | "objective_values = []" |
||
| 379 | ] |
||
| 380 | }, |
||
| 381 | { |
||
| 382 | "cell_type": "markdown", |
||
| 383 | "id": "legendary-sherman", |
||
| 384 | "metadata": {}, |
||
| 385 | "source": [ |
||
| 386 | "We re-use a callback function from the Neural Network Classifier & Regressor tutorial to plot iteration versus objective function value with some minor tweaks to plot objective values at each step." |
||
| 387 | ] |
||
| 388 | }, |
||
| 389 | { |
||
| 390 | "cell_type": "code", |
||
| 391 | "execution_count": 247, |
||
| 392 | "id": "periodic-apparel", |
||
| 393 | "metadata": {}, |
||
| 394 | "outputs": [], |
||
| 395 | "source": [ |
||
| 396 | "# callback function that draws a live plot when the .fit() method is called\n", |
||
| 397 | "def callback_graph(_, objective_value):\n", |
||
| 398 | " clear_output(wait=True)\n", |
||
| 399 | " objective_values.append(objective_value)\n", |
||
| 400 | "\n", |
||
| 401 | " plt.title(\"Objective function value against iteration\")\n", |
||
| 402 | " plt.xlabel(\"Iteration\")\n", |
||
| 403 | " plt.ylabel(\"Objective function value\")\n", |
||
| 404 | "\n", |
||
| 405 | " stage1_len = np.min((len(objective_values), maxiter))\n", |
||
| 406 | " stage1_x = np.linspace(1, stage1_len, stage1_len)\n", |
||
| 407 | " stage1_y = objective_values[:stage1_len]\n", |
||
| 408 | "\n", |
||
| 409 | " stage2_len = np.max((0, len(objective_values) - maxiter))\n", |
||
| 410 | " stage2_x = np.linspace(maxiter, maxiter + stage2_len - 1, stage2_len)\n", |
||
| 411 | " stage2_y = objective_values[maxiter : maxiter + stage2_len]\n", |
||
| 412 | "\n", |
||
| 413 | " plt.plot(stage1_x, stage1_y, color=\"orange\")\n", |
||
| 414 | " plt.plot(stage2_x, stage2_y, color=\"purple\")\n", |
||
| 415 | " plt.show()\n", |
||
| 416 | "\n", |
||
| 417 | "\n", |
||
| 418 | "plt.rcParams[\"figure.figsize\"] = (12, 6)" |
||
| 419 | ] |
||
| 420 | }, |
||
| 421 | { |
||
| 422 | "cell_type": "markdown", |
||
| 423 | "id": "institutional-cyprus", |
||
| 424 | "metadata": {}, |
||
| 425 | "source": [ |
||
| 426 | "As mentioned above we train a `VQC` model and set `COBYLA` as an optimizer with a chosen value of the `maxiter` parameter. Then we evaluate performance of the model to see how well it was trained. Then we save this model for a file. On the second step we load this model and will continue to work with it.\n", |
||
| 427 | "\n", |
||
| 428 | "Here, we manually construct an ansatz to fix an initial point where to start optimization from." |
||
| 429 | ] |
||
| 430 | }, |
||
| 431 | { |
||
| 432 | "cell_type": "code", |
||
| 433 | "execution_count": 248, |
||
| 434 | "id": "electronic-impact", |
||
| 435 | "metadata": {}, |
||
| 436 | "outputs": [], |
||
| 437 | "source": [ |
||
| 438 | "original_optimizer = AQGD(maxiter=maxiter)\n", |
||
| 439 | "\n", |
||
| 440 | "ansatz = RealAmplitudes(num_features)\n", |
||
| 441 | "initial_point = np.asarray([0.5] * ansatz.num_parameters)" |
||
| 442 | ] |
||
| 443 | }, |
||
| 444 | { |
||
| 445 | "cell_type": "markdown", |
||
| 446 | "id": "separated-classroom", |
||
| 447 | "metadata": {}, |
||
| 448 | "source": [ |
||
| 449 | "We create a model and set a sampler to the first sampler we created earlier." |
||
| 450 | ] |
||
| 451 | }, |
||
| 452 | { |
||
| 453 | "cell_type": "code", |
||
| 454 | "execution_count": 249, |
||
| 455 | "id": "revolutionary-freeze", |
||
| 456 | "metadata": {}, |
||
| 457 | "outputs": [], |
||
| 458 | "source": [ |
||
| 459 | "original_classifier = VQC(\n", |
||
| 460 | " ansatz=ansatz, optimizer=original_optimizer, callback=callback_graph, sampler=sampler1\n", |
||
| 461 | ")" |
||
| 462 | ] |
||
| 463 | }, |
||
| 464 | { |
||
| 465 | "cell_type": "markdown", |
||
| 466 | "id": "minute-mexican", |
||
| 467 | "metadata": {}, |
||
| 468 | "source": [ |
||
| 469 | "Now it is time to train the model." |
||
| 470 | ] |
||
| 471 | }, |
||
| 472 | { |
||
| 473 | "cell_type": "code", |
||
| 474 | "execution_count": 250, |
||
| 475 | "id": "suited-appointment", |
||
| 476 | "metadata": {}, |
||
| 477 | "outputs": [ |
||
| 478 | { |
||
| 479 | "data": { |
||
| 480 | "image/png": "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\n", |
||
| 481 | "text/plain": [ |
||
| 482 | "<Figure size 1200x600 with 1 Axes>" |
||
| 483 | ] |
||
| 484 | }, |
||
| 485 | "metadata": {}, |
||
| 486 | "output_type": "display_data" |
||
| 487 | }, |
||
| 488 | { |
||
| 489 | "data": { |
||
| 490 | "text/plain": [ |
||
| 491 | "<qiskit_machine_learning.algorithms.classifiers.vqc.VQC at 0x7f8e09335a80>" |
||
| 492 | ] |
||
| 493 | }, |
||
| 494 | "execution_count": 250, |
||
| 495 | "metadata": {}, |
||
| 496 | "output_type": "execute_result" |
||
| 497 | } |
||
| 498 | ], |
||
| 499 | "source": [ |
||
| 500 | "original_classifier.fit(train_features, train_labels)" |
||
| 501 | ] |
||
| 502 | }, |
||
| 503 | { |
||
| 504 | "cell_type": "markdown", |
||
| 505 | "id": "revised-torture", |
||
| 506 | "metadata": {}, |
||
| 507 | "source": [ |
||
| 508 | "Let's see how well our model performs after the first step of training." |
||
| 509 | ] |
||
| 510 | }, |
||
| 511 | { |
||
| 512 | "cell_type": "code", |
||
| 513 | "execution_count": 251, |
||
| 514 | "id": "greek-memphis", |
||
| 515 | "metadata": {}, |
||
| 516 | "outputs": [ |
||
| 517 | { |
||
| 518 | "name": "stdout", |
||
| 519 | "output_type": "stream", |
||
| 520 | "text": [ |
||
| 521 | "Train score 0.8\n", |
||
| 522 | "Test score 0.75\n" |
||
| 523 | ] |
||
| 524 | } |
||
| 525 | ], |
||
| 526 | "source": [ |
||
| 527 | "print(\"Train score\", original_classifier.score(train_features, train_labels))\n", |
||
| 528 | "print(\"Test score \", original_classifier.score(test_features, test_labels))" |
||
| 529 | ] |
||
| 530 | }, |
||
| 531 | { |
||
| 532 | "cell_type": "markdown", |
||
| 533 | "id": "rental-moses", |
||
| 534 | "metadata": {}, |
||
| 535 | "source": [ |
||
| 536 | "Next, we save the model. You may choose any file name you want. Please note that the `save` method does not append an extension if it is not specified in the file name." |
||
| 537 | ] |
||
| 538 | }, |
||
| 539 | { |
||
| 540 | "cell_type": "code", |
||
| 541 | "execution_count": 252, |
||
| 542 | "id": "broadband-interview", |
||
| 543 | "metadata": {}, |
||
| 544 | "outputs": [], |
||
| 545 | "source": [ |
||
| 546 | "original_classifier.save(\"vqc_classifier.model\")" |
||
| 547 | ] |
||
| 548 | }, |
||
| 549 | { |
||
| 550 | "cell_type": "markdown", |
||
| 551 | "id": "sitting-thread", |
||
| 552 | "metadata": {}, |
||
| 553 | "source": [ |
||
| 554 | "## 3. Load a model and continue training\n", |
||
| 555 | "\n", |
||
| 556 | "To load a model a user have to call a class method `load` of the corresponding model class. In our case it is `VQC`. We pass the same file name we used in the previous section where we saved our model." |
||
| 557 | ] |
||
| 558 | }, |
||
| 559 | { |
||
| 560 | "cell_type": "code", |
||
| 561 | "execution_count": 253, |
||
| 562 | "id": "steady-europe", |
||
| 563 | "metadata": {}, |
||
| 564 | "outputs": [], |
||
| 565 | "source": [ |
||
| 566 | "loaded_classifier = VQC.load(\"vqc_classifier.model\")" |
||
| 567 | ] |
||
| 568 | }, |
||
| 569 | { |
||
| 570 | "cell_type": "markdown", |
||
| 571 | "id": "reverse-shaft", |
||
| 572 | "metadata": {}, |
||
| 573 | "source": [ |
||
| 574 | "Next, we want to alter the model in a way it can be trained further and on another simulator. To do so, we set the `warm_start` property. When it is set to `True` and `fit()` is called again the model uses weights from previous fit to start a new fit. We also set the `sampler` property of the underlying network to the second instance of the `Sampler` primitive we created in the beginning of the tutorial. Finally, we create and set a new optimizer with `maxiter` is set to `80`, so the total number of iterations is `100`." |
||
| 575 | ] |
||
| 576 | }, |
||
| 577 | { |
||
| 578 | "cell_type": "code", |
||
| 579 | "execution_count": 254, |
||
| 580 | "id": "accessible-cowboy", |
||
| 581 | "metadata": {}, |
||
| 582 | "outputs": [], |
||
| 583 | "source": [ |
||
| 584 | "loaded_classifier.warm_start = True\n", |
||
| 585 | "loaded_classifier.neural_network.sampler = sampler2\n", |
||
| 586 | "loaded_classifier.optimizer = AQGD(maxiter=80)" |
||
| 587 | ] |
||
| 588 | }, |
||
| 589 | { |
||
| 590 | "cell_type": "markdown", |
||
| 591 | "id": "revised-bruce", |
||
| 592 | "metadata": {}, |
||
| 593 | "source": [ |
||
| 594 | "Now we continue training our model from the state we finished in the previous section." |
||
| 595 | ] |
||
| 596 | }, |
||
| 597 | { |
||
| 598 | "cell_type": "code", |
||
| 599 | "execution_count": 255, |
||
| 600 | "id": "metric-cyprus", |
||
| 601 | "metadata": { |
||
| 602 | "nbsphinx-thumbnail": { |
||
| 603 | "output-index": 0 |
||
| 604 | } |
||
| 605 | }, |
||
| 606 | "outputs": [ |
||
| 607 | { |
||
| 608 | "data": { |
||
| 609 | "image/png": "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\n", |
||
| 610 | "text/plain": [ |
||
| 611 | "<Figure size 1200x600 with 1 Axes>" |
||
| 612 | ] |
||
| 613 | }, |
||
| 614 | "metadata": {}, |
||
| 615 | "output_type": "display_data" |
||
| 616 | }, |
||
| 617 | { |
||
| 618 | "data": { |
||
| 619 | "text/plain": [ |
||
| 620 | "<qiskit_machine_learning.algorithms.classifiers.vqc.VQC at 0x7f8e08b12e90>" |
||
| 621 | ] |
||
| 622 | }, |
||
| 623 | "execution_count": 255, |
||
| 624 | "metadata": {}, |
||
| 625 | "output_type": "execute_result" |
||
| 626 | } |
||
| 627 | ], |
||
| 628 | "source": [ |
||
| 629 | "loaded_classifier.fit(train_features, train_labels)" |
||
| 630 | ] |
||
| 631 | }, |
||
| 632 | { |
||
| 633 | "cell_type": "code", |
||
| 634 | "execution_count": 256, |
||
| 635 | "id": "bronze-spread", |
||
| 636 | "metadata": {}, |
||
| 637 | "outputs": [ |
||
| 638 | { |
||
| 639 | "name": "stdout", |
||
| 640 | "output_type": "stream", |
||
| 641 | "text": [ |
||
| 642 | "Train score 0.8666666666666667\n", |
||
| 643 | "Test score 0.9\n" |
||
| 644 | ] |
||
| 645 | } |
||
| 646 | ], |
||
| 647 | "source": [ |
||
| 648 | "print(\"Train score\", loaded_classifier.score(train_features, train_labels))\n", |
||
| 649 | "print(\"Test score\", loaded_classifier.score(test_features, test_labels))" |
||
| 650 | ] |
||
| 651 | }, |
||
| 652 | { |
||
| 653 | "cell_type": "markdown", |
||
| 654 | "id": "apparent-bloom", |
||
| 655 | "metadata": {}, |
||
| 656 | "source": [ |
||
| 657 | "Let's see which data points were misclassified. First, we call `predict` to infer predicted values from the training and test features." |
||
| 658 | ] |
||
| 659 | }, |
||
| 660 | { |
||
| 661 | "cell_type": "code", |
||
| 662 | "execution_count": 257, |
||
| 663 | "id": "catholic-norway", |
||
| 664 | "metadata": {}, |
||
| 665 | "outputs": [], |
||
| 666 | "source": [ |
||
| 667 | "train_predicts = loaded_classifier.predict(train_features)\n", |
||
| 668 | "test_predicts = loaded_classifier.predict(test_features)" |
||
| 669 | ] |
||
| 670 | }, |
||
| 671 | { |
||
| 672 | "cell_type": "markdown", |
||
| 673 | "id": "guided-croatia", |
||
| 674 | "metadata": {}, |
||
| 675 | "source": [ |
||
| 676 | "Plot the whole dataset and the highlight the points that were classified incorrectly." |
||
| 677 | ] |
||
| 678 | }, |
||
| 679 | { |
||
| 680 | "cell_type": "code", |
||
| 681 | "execution_count": 258, |
||
| 682 | "id": "tested-handling", |
||
| 683 | "metadata": {}, |
||
| 684 | "outputs": [ |
||
| 685 | { |
||
| 686 | "data": { |
||
| 687 | "text/plain": [ |
||
| 688 | "<matplotlib.collections.PathCollection at 0x7f8e443682e0>" |
||
| 689 | ] |
||
| 690 | }, |
||
| 691 | "execution_count": 258, |
||
| 692 | "metadata": {}, |
||
| 693 | "output_type": "execute_result" |
||
| 694 | }, |
||
| 695 | { |
||
| 696 | "data": { |
||
| 697 | "image/png": "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\n", |
||
| 698 | "text/plain": [ |
||
| 699 | "<Figure size 600x400 with 1 Axes>" |
||
| 700 | ] |
||
| 701 | }, |
||
| 702 | "metadata": {}, |
||
| 703 | "output_type": "display_data" |
||
| 704 | } |
||
| 705 | ], |
||
| 706 | "source": [ |
||
| 707 | "# return plot to default figsize\n", |
||
| 708 | "plt.rcParams[\"figure.figsize\"] = (6, 4)\n", |
||
| 709 | "\n", |
||
| 710 | "plot_dataset()\n", |
||
| 711 | "\n", |
||
| 712 | "# plot misclassified data points\n", |
||
| 713 | "plt.scatter(\n", |
||
| 714 | " train_features[np.all(train_labels != train_predicts, axis=1), 0],\n", |
||
| 715 | " train_features[np.all(train_labels != train_predicts, axis=1), 1],\n", |
||
| 716 | " s=200,\n", |
||
| 717 | " facecolors=\"none\",\n", |
||
| 718 | " edgecolors=\"r\",\n", |
||
| 719 | " linewidths=2,\n", |
||
| 720 | ")\n", |
||
| 721 | "plt.scatter(\n", |
||
| 722 | " test_features[np.all(test_labels != test_predicts, axis=1), 0],\n", |
||
| 723 | " test_features[np.all(test_labels != test_predicts, axis=1), 1],\n", |
||
| 724 | " s=200,\n", |
||
| 725 | " facecolors=\"none\",\n", |
||
| 726 | " edgecolors=\"r\",\n", |
||
| 727 | " linewidths=2,\n", |
||
| 728 | ")" |
||
| 729 | ] |
||
| 730 | }, |
||
| 731 | { |
||
| 732 | "cell_type": "markdown", |
||
| 733 | "id": "genuine-preference", |
||
| 734 | "metadata": {}, |
||
| 735 | "source": [ |
||
| 736 | "So, if you have a large dataset or a large model you can train it in multiple steps as shown in this tutorial." |
||
| 737 | ] |
||
| 738 | }, |
||
| 739 | { |
||
| 740 | "cell_type": "markdown", |
||
| 741 | "id": "acknowledged-freight", |
||
| 742 | "metadata": {}, |
||
| 743 | "source": [ |
||
| 744 | "## 4. PyTorch hybrid models\n", |
||
| 745 | "\n", |
||
| 746 | "To save and load hybrid models, when using the TorchConnector, follow the PyTorch recommendations of saving and loading the models. For more details please refer to the [PyTorch Connector tutorial](05_torch_connector.ipynb) where a short snippet shows how to do it.\n", |
||
| 747 | "\n", |
||
| 748 | "Take a look at this pseudo-like code to get the idea:\n", |
||
| 749 | "```python\n", |
||
| 750 | "# create a QNN and a hybrid model\n", |
||
| 751 | "qnn = create_qnn()\n", |
||
| 752 | "model = Net(qnn)\n", |
||
| 753 | "# ... train the model ...\n", |
||
| 754 | "\n", |
||
| 755 | "# save the model\n", |
||
| 756 | "torch.save(model.state_dict(), \"model.pt\")\n", |
||
| 757 | "\n", |
||
| 758 | "# create a new model\n", |
||
| 759 | "new_qnn = create_qnn()\n", |
||
| 760 | "loaded_model = Net(new_qnn)\n", |
||
| 761 | "loaded_model.load_state_dict(torch.load(\"model.pt\"))\n", |
||
| 762 | "```" |
||
| 763 | ] |
||
| 764 | }, |
||
| 765 | { |
||
| 766 | "cell_type": "code", |
||
| 767 | "execution_count": 259, |
||
| 768 | "id": "persistent-combine", |
||
| 769 | "metadata": { |
||
| 770 | "tags": [] |
||
| 771 | }, |
||
| 772 | "outputs": [ |
||
| 773 | { |
||
| 774 | "data": { |
||
| 775 | "text/html": [ |
||
| 776 | "<h3>Version Information</h3><table><tr><th>Software</th><th>Version</th></tr><tr><td><code>qiskit</code></td><td>0.44.1</td></tr><tr><td><code>qiskit-terra</code></td><td>0.25.1</td></tr><tr><td><code>qiskit_machine_learning</code></td><td>0.6.1</td></tr><tr><th colspan='2'>System information</th></tr><tr><td>Python version</td><td>3.10.8</td></tr><tr><td>Python compiler</td><td>GCC 10.4.0</td></tr><tr><td>Python build</td><td>main, Nov 22 2022 08:26:04</td></tr><tr><td>OS</td><td>Linux</td></tr><tr><td>CPUs</td><td>8</td></tr><tr><td>Memory (Gb)</td><td>31.142810821533203</td></tr><tr><td colspan='2'>Wed Nov 22 10:18:49 2023 UTC</td></tr></table>" |
||
| 777 | ], |
||
| 778 | "text/plain": [ |
||
| 779 | "<IPython.core.display.HTML object>" |
||
| 780 | ] |
||
| 781 | }, |
||
| 782 | "metadata": {}, |
||
| 783 | "output_type": "display_data" |
||
| 784 | }, |
||
| 785 | { |
||
| 786 | "data": { |
||
| 787 | "text/html": [ |
||
| 788 | "<div style='width: 100%; background-color:#d5d9e0;padding-left: 10px; padding-bottom: 10px; padding-right: 10px; padding-top: 5px'><h3>This code is a part of Qiskit</h3><p>© Copyright IBM 2017, 2023.</p><p>This code is licensed under the Apache License, Version 2.0. You may<br>obtain a copy of this license in the LICENSE.txt file in the root directory<br> of this source tree or at http://www.apache.org/licenses/LICENSE-2.0.<p>Any modifications or derivative works of this code must retain this<br>copyright notice, and modified files need to carry a notice indicating<br>that they have been altered from the originals.</p></div>" |
||
| 789 | ], |
||
| 790 | "text/plain": [ |
||
| 791 | "<IPython.core.display.HTML object>" |
||
| 792 | ] |
||
| 793 | }, |
||
| 794 | "metadata": {}, |
||
| 795 | "output_type": "display_data" |
||
| 796 | } |
||
| 797 | ], |
||
| 798 | "source": [ |
||
| 799 | "import qiskit.tools.jupyter\n", |
||
| 800 | "\n", |
||
| 801 | "%qiskit_version_table\n", |
||
| 802 | "%qiskit_copyright" |
||
| 803 | ] |
||
| 804 | } |
||
| 805 | ], |
||
| 806 | "metadata": { |
||
| 807 | "celltoolbar": "Tags", |
||
| 808 | "kernelspec": { |
||
| 809 | "display_name": "Python 3 (ipykernel)", |
||
| 810 | "language": "python", |
||
| 811 | "name": "python3" |
||
| 812 | }, |
||
| 813 | "language_info": { |
||
| 814 | "codemirror_mode": { |
||
| 815 | "name": "ipython", |
||
| 816 | "version": 3 |
||
| 817 | }, |
||
| 818 | "file_extension": ".py", |
||
| 819 | "mimetype": "text/x-python", |
||
| 820 | "name": "python", |
||
| 821 | "nbconvert_exporter": "python", |
||
| 822 | "pygments_lexer": "ipython3", |
||
| 823 | "version": "3.10.8" |
||
| 824 | }, |
||
| 825 | "widgets": { |
||
| 826 | "application/vnd.jupyter.widget-state+json": { |
||
| 827 | "state": { |
||
| 828 | "03a293b6a5874d15b6048d8c1a669a5f": { |
||
| 829 | "model_module": "@jupyter-widgets/base", |
||
| 830 | "model_module_version": "2.0.0", |
||
| 831 | "model_name": "LayoutModel", |
||
| 832 | "state": { |
||
| 833 | "width": "70px" |
||
| 834 | } |
||
| 835 | }, |
||
| 836 | "0ad579d545d446fb8326902b61fe7ef6": { |
||
| 837 | "model_module": "@jupyter-widgets/controls", |
||
| 838 | "model_module_version": "2.0.0", |
||
| 839 | "model_name": "ButtonModel", |
||
| 840 | "state": { |
||
| 841 | "button_style": "primary", |
||
| 842 | "description": "Clear", |
||
| 843 | "layout": "IPY_MODEL_b4763ce700694273b12f43e85b337363", |
||
| 844 | "style": "IPY_MODEL_59a150a8431d41e9831736f3fe176599", |
||
| 845 | "tooltip": null |
||
| 846 | } |
||
| 847 | }, |
||
| 848 | "2d29cdbbfd8146d49ad4fb5077f71dab": { |
||
| 849 | "model_module": "@jupyter-widgets/controls", |
||
| 850 | "model_module_version": "2.0.0", |
||
| 851 | "model_name": "HTMLStyleModel", |
||
| 852 | "state": { |
||
| 853 | "description_width": "", |
||
| 854 | "font_size": null, |
||
| 855 | "text_color": null |
||
| 856 | } |
||
| 857 | }, |
||
| 858 | "50a5d0cb87274ce68ec3218a139f9784": { |
||
| 859 | "model_module": "@jupyter-widgets/base", |
||
| 860 | "model_module_version": "2.0.0", |
||
| 861 | "model_name": "LayoutModel", |
||
| 862 | "state": { |
||
| 863 | "margin": "0px 0px 10px 0px" |
||
| 864 | } |
||
| 865 | }, |
||
| 866 | "52e9cd11d14c42a6bb1be82655505e8a": { |
||
| 867 | "model_module": "@jupyter-widgets/controls", |
||
| 868 | "model_module_version": "2.0.0", |
||
| 869 | "model_name": "HTMLModel", |
||
| 870 | "state": { |
||
| 871 | "layout": "IPY_MODEL_60bc550a402840a597b294aeab2a690f", |
||
| 872 | "style": "IPY_MODEL_2d29cdbbfd8146d49ad4fb5077f71dab", |
||
| 873 | "value": "<h5>Status</h5>" |
||
| 874 | } |
||
| 875 | }, |
||
| 876 | "59a150a8431d41e9831736f3fe176599": { |
||
| 877 | "model_module": "@jupyter-widgets/controls", |
||
| 878 | "model_module_version": "2.0.0", |
||
| 879 | "model_name": "ButtonStyleModel", |
||
| 880 | "state": { |
||
| 881 | "font_family": null, |
||
| 882 | "font_size": null, |
||
| 883 | "font_style": null, |
||
| 884 | "font_variant": null, |
||
| 885 | "font_weight": null, |
||
| 886 | "text_color": null, |
||
| 887 | "text_decoration": null |
||
| 888 | } |
||
| 889 | }, |
||
| 890 | "5c7b8a96279a4ec7be226685f3bd46c5": { |
||
| 891 | "model_module": "@jupyter-widgets/base", |
||
| 892 | "model_module_version": "2.0.0", |
||
| 893 | "model_name": "LayoutModel", |
||
| 894 | "state": { |
||
| 895 | "grid_template_areas": "\n \". . . . right \"\n ", |
||
| 896 | "grid_template_columns": "20% 20% 20% 20% 20%", |
||
| 897 | "width": "100%" |
||
| 898 | } |
||
| 899 | }, |
||
| 900 | "5d7f1676158a4c8aaae790dd4c00851b": { |
||
| 901 | "model_module": "@jupyter-widgets/base", |
||
| 902 | "model_module_version": "2.0.0", |
||
| 903 | "model_name": "LayoutModel", |
||
| 904 | "state": { |
||
| 905 | "margin": "0px 0px 0px 37px", |
||
| 906 | "width": "600px" |
||
| 907 | } |
||
| 908 | }, |
||
| 909 | "60bc550a402840a597b294aeab2a690f": { |
||
| 910 | "model_module": "@jupyter-widgets/base", |
||
| 911 | "model_module_version": "2.0.0", |
||
| 912 | "model_name": "LayoutModel", |
||
| 913 | "state": { |
||
| 914 | "width": "95px" |
||
| 915 | } |
||
| 916 | }, |
||
| 917 | "663d495d2a4e4595a8637bb374a4d04f": { |
||
| 918 | "model_module": "@jupyter-widgets/controls", |
||
| 919 | "model_module_version": "2.0.0", |
||
| 920 | "model_name": "HTMLStyleModel", |
||
| 921 | "state": { |
||
| 922 | "description_width": "", |
||
| 923 | "font_size": null, |
||
| 924 | "text_color": null |
||
| 925 | } |
||
| 926 | }, |
||
| 927 | "8efaa0d11d204d359c6b2f52d27ce586": { |
||
| 928 | "model_module": "@jupyter-widgets/controls", |
||
| 929 | "model_module_version": "2.0.0", |
||
| 930 | "model_name": "HTMLModel", |
||
| 931 | "state": { |
||
| 932 | "layout": "IPY_MODEL_ca6772bfdb2148c4a9a56eba40716259", |
||
| 933 | "style": "IPY_MODEL_f18130f8e0624484ae3a9268f502d20c", |
||
| 934 | "value": "<h5>Message</h5>" |
||
| 935 | } |
||
| 936 | }, |
||
| 937 | "90939a5b2766492ea9483fdb20558bd8": { |
||
| 938 | "model_module": "@jupyter-widgets/controls", |
||
| 939 | "model_module_version": "2.0.0", |
||
| 940 | "model_name": "HTMLStyleModel", |
||
| 941 | "state": { |
||
| 942 | "description_width": "", |
||
| 943 | "font_size": null, |
||
| 944 | "text_color": null |
||
| 945 | } |
||
| 946 | }, |
||
| 947 | "a87c48aad3fb44769246d8556db63000": { |
||
| 948 | "model_module": "@jupyter-widgets/controls", |
||
| 949 | "model_module_version": "2.0.0", |
||
| 950 | "model_name": "GridBoxModel", |
||
| 951 | "state": { |
||
| 952 | "children": [ |
||
| 953 | "IPY_MODEL_0ad579d545d446fb8326902b61fe7ef6" |
||
| 954 | ], |
||
| 955 | "layout": "IPY_MODEL_5c7b8a96279a4ec7be226685f3bd46c5" |
||
| 956 | } |
||
| 957 | }, |
||
| 958 | "b1c84eb3398a43549f837b604ec39d83": { |
||
| 959 | "model_module": "@jupyter-widgets/controls", |
||
| 960 | "model_module_version": "2.0.0", |
||
| 961 | "model_name": "HTMLModel", |
||
| 962 | "state": { |
||
| 963 | "layout": "IPY_MODEL_50a5d0cb87274ce68ec3218a139f9784", |
||
| 964 | "style": "IPY_MODEL_90939a5b2766492ea9483fdb20558bd8", |
||
| 965 | "value": "<p style='font-family: IBM Plex Sans, Arial, Helvetica, sans-serif; font-size: 20px; font-weight: medium;'>Circuit Properties</p>" |
||
| 966 | } |
||
| 967 | }, |
||
| 968 | "b28fa4fc45c74a1281f8c1cff661b8c3": { |
||
| 969 | "model_module": "@jupyter-widgets/controls", |
||
| 970 | "model_module_version": "2.0.0", |
||
| 971 | "model_name": "HTMLModel", |
||
| 972 | "state": { |
||
| 973 | "layout": "IPY_MODEL_d5c8577291cc461386fa37e051d3c647", |
||
| 974 | "style": "IPY_MODEL_663d495d2a4e4595a8637bb374a4d04f", |
||
| 975 | "value": "<h5>Backend</h5>" |
||
| 976 | } |
||
| 977 | }, |
||
| 978 | "b4763ce700694273b12f43e85b337363": { |
||
| 979 | "model_module": "@jupyter-widgets/base", |
||
| 980 | "model_module_version": "2.0.0", |
||
| 981 | "model_name": "LayoutModel", |
||
| 982 | "state": { |
||
| 983 | "grid_area": "right", |
||
| 984 | "padding": "0px 0px 0px 0px", |
||
| 985 | "width": "70px" |
||
| 986 | } |
||
| 987 | }, |
||
| 988 | "b8f1e7a508dc42bc912feb4594b7311d": { |
||
| 989 | "model_module": "@jupyter-widgets/controls", |
||
| 990 | "model_module_version": "2.0.0", |
||
| 991 | "model_name": "HTMLStyleModel", |
||
| 992 | "state": { |
||
| 993 | "description_width": "", |
||
| 994 | "font_size": null, |
||
| 995 | "text_color": null |
||
| 996 | } |
||
| 997 | }, |
||
| 998 | "beb929c105fb4464bc41ad0be74939dc": { |
||
| 999 | "model_module": "@jupyter-widgets/controls", |
||
| 1000 | "model_module_version": "2.0.0", |
||
| 1001 | "model_name": "HTMLStyleModel", |
||
| 1002 | "state": { |
||
| 1003 | "description_width": "", |
||
| 1004 | "font_size": null, |
||
| 1005 | "text_color": null |
||
| 1006 | } |
||
| 1007 | }, |
||
| 1008 | "ca6772bfdb2148c4a9a56eba40716259": { |
||
| 1009 | "model_module": "@jupyter-widgets/base", |
||
| 1010 | "model_module_version": "2.0.0", |
||
| 1011 | "model_name": "LayoutModel", |
||
| 1012 | "state": {} |
||
| 1013 | }, |
||
| 1014 | "d5c8577291cc461386fa37e051d3c647": { |
||
| 1015 | "model_module": "@jupyter-widgets/base", |
||
| 1016 | "model_module_version": "2.0.0", |
||
| 1017 | "model_name": "LayoutModel", |
||
| 1018 | "state": { |
||
| 1019 | "width": "145px" |
||
| 1020 | } |
||
| 1021 | }, |
||
| 1022 | "d752d39c18fc44589713b7c4dd284cfc": { |
||
| 1023 | "model_module": "@jupyter-widgets/base", |
||
| 1024 | "model_module_version": "2.0.0", |
||
| 1025 | "model_name": "LayoutModel", |
||
| 1026 | "state": { |
||
| 1027 | "width": "190px" |
||
| 1028 | } |
||
| 1029 | }, |
||
| 1030 | "f12aa440028144bc8ce098d59d19ef20": { |
||
| 1031 | "model_module": "@jupyter-widgets/controls", |
||
| 1032 | "model_module_version": "2.0.0", |
||
| 1033 | "model_name": "HBoxModel", |
||
| 1034 | "state": { |
||
| 1035 | "children": [ |
||
| 1036 | "IPY_MODEL_fe483ef8a25d43b58f42f6b27a4107b1", |
||
| 1037 | "IPY_MODEL_b28fa4fc45c74a1281f8c1cff661b8c3", |
||
| 1038 | "IPY_MODEL_52e9cd11d14c42a6bb1be82655505e8a", |
||
| 1039 | "IPY_MODEL_f690148eb1d94330bd301caf59e0c0b5", |
||
| 1040 | "IPY_MODEL_8efaa0d11d204d359c6b2f52d27ce586" |
||
| 1041 | ], |
||
| 1042 | "layout": "IPY_MODEL_5d7f1676158a4c8aaae790dd4c00851b" |
||
| 1043 | } |
||
| 1044 | }, |
||
| 1045 | "f18130f8e0624484ae3a9268f502d20c": { |
||
| 1046 | "model_module": "@jupyter-widgets/controls", |
||
| 1047 | "model_module_version": "2.0.0", |
||
| 1048 | "model_name": "HTMLStyleModel", |
||
| 1049 | "state": { |
||
| 1050 | "description_width": "", |
||
| 1051 | "font_size": null, |
||
| 1052 | "text_color": null |
||
| 1053 | } |
||
| 1054 | }, |
||
| 1055 | "f690148eb1d94330bd301caf59e0c0b5": { |
||
| 1056 | "model_module": "@jupyter-widgets/controls", |
||
| 1057 | "model_module_version": "2.0.0", |
||
| 1058 | "model_name": "HTMLModel", |
||
| 1059 | "state": { |
||
| 1060 | "layout": "IPY_MODEL_03a293b6a5874d15b6048d8c1a669a5f", |
||
| 1061 | "style": "IPY_MODEL_beb929c105fb4464bc41ad0be74939dc", |
||
| 1062 | "value": "<h5>Queue</h5>" |
||
| 1063 | } |
||
| 1064 | }, |
||
| 1065 | "fe483ef8a25d43b58f42f6b27a4107b1": { |
||
| 1066 | "model_module": "@jupyter-widgets/controls", |
||
| 1067 | "model_module_version": "2.0.0", |
||
| 1068 | "model_name": "HTMLModel", |
||
| 1069 | "state": { |
||
| 1070 | "layout": "IPY_MODEL_d752d39c18fc44589713b7c4dd284cfc", |
||
| 1071 | "style": "IPY_MODEL_b8f1e7a508dc42bc912feb4594b7311d", |
||
| 1072 | "value": "<h5>Job ID</h5>" |
||
| 1073 | } |
||
| 1074 | } |
||
| 1075 | }, |
||
| 1076 | "version_major": 2, |
||
| 1077 | "version_minor": 0 |
||
| 1078 | } |
||
| 1079 | } |
||
| 1080 | }, |
||
| 1081 | "nbformat": 4, |
||
| 1082 | "nbformat_minor": 5 |
||
| 1083 | } |