Commit 17772b9cbd7814531c63f265c3552032fea551a4
1 parent
d290a4d480
Exists in
main
Update test_mpg.ipynb
Showing 1 changed file with 597 additions and 19 deletions Side-by-side Diff
python-notebook/test_mpg.ipynb
View file @
17772b9
| ... | ... | @@ -2,7 +2,7 @@ |
| 2 | 2 | "cells": [ |
| 3 | 3 | { |
| 4 | 4 | "cell_type": "code", |
| 5 | - "execution_count": 1, | |
| 5 | + "execution_count": 18, | |
| 6 | 6 | "metadata": {}, |
| 7 | 7 | "outputs": [ |
| 8 | 8 | { |
| 9 | 9 | |
| 10 | 10 | |
| ... | ... | @@ -29,25 +29,16 @@ |
| 29 | 29 | }, |
| 30 | 30 | { |
| 31 | 31 | "cell_type": "code", |
| 32 | - "execution_count": 2, | |
| 32 | + "execution_count": 19, | |
| 33 | 33 | "metadata": {}, |
| 34 | 34 | "outputs": [ |
| 35 | 35 | { |
| 36 | - "name": "stdout", | |
| 37 | - "output_type": "stream", | |
| 38 | - "text": [ | |
| 39 | - "Downloading data from http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data\n", | |
| 40 | - "32768/30286 [================================] - 0s 1us/step\n", | |
| 41 | - "40960/30286 [========================================] - 0s 1us/step\n" | |
| 42 | - ] | |
| 43 | - }, | |
| 44 | - { | |
| 45 | 36 | "data": { |
| 46 | 37 | "text/plain": [ |
| 47 | 38 | "'/Users/ffee21/.keras/datasets/auto-mpg.data'" |
| 48 | 39 | ] |
| 49 | 40 | }, |
| 50 | - "execution_count": 2, | |
| 41 | + "execution_count": 19, | |
| 51 | 42 | "metadata": {}, |
| 52 | 43 | "output_type": "execute_result" |
| 53 | 44 | } |
| ... | ... | @@ -59,7 +50,7 @@ |
| 59 | 50 | }, |
| 60 | 51 | { |
| 61 | 52 | "cell_type": "code", |
| 62 | - "execution_count": 12, | |
| 53 | + "execution_count": 20, | |
| 63 | 54 | "metadata": {}, |
| 64 | 55 | "outputs": [ |
| 65 | 56 | { |
| ... | ... | @@ -181,7 +172,7 @@ |
| 181 | 172 | "397 82 1.0 0.0 0.0 " |
| 182 | 173 | ] |
| 183 | 174 | }, |
| 184 | - "execution_count": 12, | |
| 175 | + "execution_count": 20, | |
| 185 | 176 | "metadata": {}, |
| 186 | 177 | "output_type": "execute_result" |
| 187 | 178 | } |
| ... | ... | @@ -206,7 +197,7 @@ |
| 206 | 197 | }, |
| 207 | 198 | { |
| 208 | 199 | "cell_type": "code", |
| 209 | - "execution_count": 13, | |
| 200 | + "execution_count": 21, | |
| 210 | 201 | "metadata": {}, |
| 211 | 202 | "outputs": [], |
| 212 | 203 | "source": [ |
| 213 | 204 | |
| 214 | 205 | |
| 215 | 206 | |
| ... | ... | @@ -216,16 +207,169 @@ |
| 216 | 207 | }, |
| 217 | 208 | { |
| 218 | 209 | "cell_type": "code", |
| 219 | - "execution_count": 14, | |
| 210 | + "execution_count": 22, | |
| 220 | 211 | "metadata": {}, |
| 221 | 212 | "outputs": [ |
| 222 | 213 | { |
| 223 | 214 | "data": { |
| 215 | + "text/html": [ | |
| 216 | + "<div>\n", | |
| 217 | + "<style scoped>\n", | |
| 218 | + " .dataframe tbody tr th:only-of-type {\n", | |
| 219 | + " vertical-align: middle;\n", | |
| 220 | + " }\n", | |
| 221 | + "\n", | |
| 222 | + " .dataframe tbody tr th {\n", | |
| 223 | + " vertical-align: top;\n", | |
| 224 | + " }\n", | |
| 225 | + "\n", | |
| 226 | + " .dataframe thead th {\n", | |
| 227 | + " text-align: right;\n", | |
| 228 | + " }\n", | |
| 229 | + "</style>\n", | |
| 230 | + "<table border=\"1\" class=\"dataframe\">\n", | |
| 231 | + " <thead>\n", | |
| 232 | + " <tr style=\"text-align: right;\">\n", | |
| 233 | + " <th></th>\n", | |
| 234 | + " <th>count</th>\n", | |
| 235 | + " <th>mean</th>\n", | |
| 236 | + " <th>std</th>\n", | |
| 237 | + " <th>min</th>\n", | |
| 238 | + " <th>25%</th>\n", | |
| 239 | + " <th>50%</th>\n", | |
| 240 | + " <th>75%</th>\n", | |
| 241 | + " <th>max</th>\n", | |
| 242 | + " </tr>\n", | |
| 243 | + " </thead>\n", | |
| 244 | + " <tbody>\n", | |
| 245 | + " <tr>\n", | |
| 246 | + " <th>Cylinders</th>\n", | |
| 247 | + " <td>314.0</td>\n", | |
| 248 | + " <td>5.477707</td>\n", | |
| 249 | + " <td>1.699788</td>\n", | |
| 250 | + " <td>3.0</td>\n", | |
| 251 | + " <td>4.00</td>\n", | |
| 252 | + " <td>4.0</td>\n", | |
| 253 | + " <td>8.00</td>\n", | |
| 254 | + " <td>8.0</td>\n", | |
| 255 | + " </tr>\n", | |
| 256 | + " <tr>\n", | |
| 257 | + " <th>Displacement</th>\n", | |
| 258 | + " <td>314.0</td>\n", | |
| 259 | + " <td>195.318471</td>\n", | |
| 260 | + " <td>104.331589</td>\n", | |
| 261 | + " <td>68.0</td>\n", | |
| 262 | + " <td>105.50</td>\n", | |
| 263 | + " <td>151.0</td>\n", | |
| 264 | + " <td>265.75</td>\n", | |
| 265 | + " <td>455.0</td>\n", | |
| 266 | + " </tr>\n", | |
| 267 | + " <tr>\n", | |
| 268 | + " <th>Horsepower</th>\n", | |
| 269 | + " <td>314.0</td>\n", | |
| 270 | + " <td>104.869427</td>\n", | |
| 271 | + " <td>38.096214</td>\n", | |
| 272 | + " <td>46.0</td>\n", | |
| 273 | + " <td>76.25</td>\n", | |
| 274 | + " <td>94.5</td>\n", | |
| 275 | + " <td>128.00</td>\n", | |
| 276 | + " <td>225.0</td>\n", | |
| 277 | + " </tr>\n", | |
| 278 | + " <tr>\n", | |
| 279 | + " <th>Weight</th>\n", | |
| 280 | + " <td>314.0</td>\n", | |
| 281 | + " <td>2990.251592</td>\n", | |
| 282 | + " <td>843.898596</td>\n", | |
| 283 | + " <td>1649.0</td>\n", | |
| 284 | + " <td>2256.50</td>\n", | |
| 285 | + " <td>2822.5</td>\n", | |
| 286 | + " <td>3608.00</td>\n", | |
| 287 | + " <td>5140.0</td>\n", | |
| 288 | + " </tr>\n", | |
| 289 | + " <tr>\n", | |
| 290 | + " <th>Acceleration</th>\n", | |
| 291 | + " <td>314.0</td>\n", | |
| 292 | + " <td>15.559236</td>\n", | |
| 293 | + " <td>2.789230</td>\n", | |
| 294 | + " <td>8.0</td>\n", | |
| 295 | + " <td>13.80</td>\n", | |
| 296 | + " <td>15.5</td>\n", | |
| 297 | + " <td>17.20</td>\n", | |
| 298 | + " <td>24.8</td>\n", | |
| 299 | + " </tr>\n", | |
| 300 | + " <tr>\n", | |
| 301 | + " <th>Model Year</th>\n", | |
| 302 | + " <td>314.0</td>\n", | |
| 303 | + " <td>75.898089</td>\n", | |
| 304 | + " <td>3.675642</td>\n", | |
| 305 | + " <td>70.0</td>\n", | |
| 306 | + " <td>73.00</td>\n", | |
| 307 | + " <td>76.0</td>\n", | |
| 308 | + " <td>79.00</td>\n", | |
| 309 | + " <td>82.0</td>\n", | |
| 310 | + " </tr>\n", | |
| 311 | + " <tr>\n", | |
| 312 | + " <th>USA</th>\n", | |
| 313 | + " <td>314.0</td>\n", | |
| 314 | + " <td>0.624204</td>\n", | |
| 315 | + " <td>0.485101</td>\n", | |
| 316 | + " <td>0.0</td>\n", | |
| 317 | + " <td>0.00</td>\n", | |
| 318 | + " <td>1.0</td>\n", | |
| 319 | + " <td>1.00</td>\n", | |
| 320 | + " <td>1.0</td>\n", | |
| 321 | + " </tr>\n", | |
| 322 | + " <tr>\n", | |
| 323 | + " <th>Europe</th>\n", | |
| 324 | + " <td>314.0</td>\n", | |
| 325 | + " <td>0.178344</td>\n", | |
| 326 | + " <td>0.383413</td>\n", | |
| 327 | + " <td>0.0</td>\n", | |
| 328 | + " <td>0.00</td>\n", | |
| 329 | + " <td>0.0</td>\n", | |
| 330 | + " <td>0.00</td>\n", | |
| 331 | + " <td>1.0</td>\n", | |
| 332 | + " </tr>\n", | |
| 333 | + " <tr>\n", | |
| 334 | + " <th>Japan</th>\n", | |
| 335 | + " <td>314.0</td>\n", | |
| 336 | + " <td>0.197452</td>\n", | |
| 337 | + " <td>0.398712</td>\n", | |
| 338 | + " <td>0.0</td>\n", | |
| 339 | + " <td>0.00</td>\n", | |
| 340 | + " <td>0.0</td>\n", | |
| 341 | + " <td>0.00</td>\n", | |
| 342 | + " <td>1.0</td>\n", | |
| 343 | + " </tr>\n", | |
| 344 | + " </tbody>\n", | |
| 345 | + "</table>\n", | |
| 346 | + "</div>" | |
| 347 | + ], | |
| 224 | 348 | "text/plain": [ |
| 225 | - "<seaborn.axisgrid.PairGrid at 0x7feb4eef3f10>" | |
| 349 | + " count mean std min 25% 50% \\\n", | |
| 350 | + "Cylinders 314.0 5.477707 1.699788 3.0 4.00 4.0 \n", | |
| 351 | + "Displacement 314.0 195.318471 104.331589 68.0 105.50 151.0 \n", | |
| 352 | + "Horsepower 314.0 104.869427 38.096214 46.0 76.25 94.5 \n", | |
| 353 | + "Weight 314.0 2990.251592 843.898596 1649.0 2256.50 2822.5 \n", | |
| 354 | + "Acceleration 314.0 15.559236 2.789230 8.0 13.80 15.5 \n", | |
| 355 | + "Model Year 314.0 75.898089 3.675642 70.0 73.00 76.0 \n", | |
| 356 | + "USA 314.0 0.624204 0.485101 0.0 0.00 1.0 \n", | |
| 357 | + "Europe 314.0 0.178344 0.383413 0.0 0.00 0.0 \n", | |
| 358 | + "Japan 314.0 0.197452 0.398712 0.0 0.00 0.0 \n", | |
| 359 | + "\n", | |
| 360 | + " 75% max \n", | |
| 361 | + "Cylinders 8.00 8.0 \n", | |
| 362 | + "Displacement 265.75 455.0 \n", | |
| 363 | + "Horsepower 128.00 225.0 \n", | |
| 364 | + "Weight 3608.00 5140.0 \n", | |
| 365 | + "Acceleration 17.20 24.8 \n", | |
| 366 | + "Model Year 79.00 82.0 \n", | |
| 367 | + "USA 1.00 1.0 \n", | |
| 368 | + "Europe 0.00 1.0 \n", | |
| 369 | + "Japan 0.00 1.0 " | |
| 226 | 370 | ] |
| 227 | 371 | }, |
| 228 | - "execution_count": 14, | |
| 372 | + "execution_count": 22, | |
| 229 | 373 | "metadata": {}, |
| 230 | 374 | "output_type": "execute_result" |
| 231 | 375 | }, |
| ... | ... | @@ -243,7 +387,441 @@ |
| 243 | 387 | } |
| 244 | 388 | ], |
| 245 | 389 | "source": [ |
| 246 | - "sns.pairplot(train_dataset[[\"MPG\", \"Cylinders\", \"Displacement\", \"Weight\"]], diag_kind=\"kde\")" | |
| 390 | + "sns.pairplot(train_dataset[[\"MPG\", \"Cylinders\", \"Displacement\", \"Weight\"]], diag_kind=\"kde\")\n", | |
| 391 | + "\n", | |
| 392 | + "train_stats = train_dataset.describe()\n", | |
| 393 | + "train_stats.pop(\"MPG\")\n", | |
| 394 | + "train_stats = train_stats.transpose()\n", | |
| 395 | + "train_stats" | |
| 396 | + ] | |
| 397 | + }, | |
| 398 | + { | |
| 399 | + "cell_type": "code", | |
| 400 | + "execution_count": 23, | |
| 401 | + "metadata": {}, | |
| 402 | + "outputs": [], | |
| 403 | + "source": [ | |
| 404 | + "train_labels = train_dataset.pop('MPG')\n", | |
| 405 | + "test_labels = test_dataset.pop('MPG')\n", | |
| 406 | + "\n", | |
| 407 | + "def norm(x):\n", | |
| 408 | + " return (x - train_stats['mean']) / train_stats['std']\n", | |
| 409 | + "normed_train_data = norm(train_dataset)\n", | |
| 410 | + "normed_test_data = norm(test_dataset)" | |
| 411 | + ] | |
| 412 | + }, | |
| 413 | + { | |
| 414 | + "cell_type": "code", | |
| 415 | + "execution_count": 24, | |
| 416 | + "metadata": {}, | |
| 417 | + "outputs": [], | |
| 418 | + "source": [ | |
| 419 | + "def build_model():\n", | |
| 420 | + " model = keras.Sequential([\n", | |
| 421 | + " layers.Dense(64, activation='relu', input_shape=[len(train_dataset.keys())]),\n", | |
| 422 | + " layers.Dense(64, activation='relu'),\n", | |
| 423 | + " layers.Dense(1)\n", | |
| 424 | + " ])\n", | |
| 425 | + "\n", | |
| 426 | + " optimizer = tf.keras.optimizers.RMSprop(0.001)\n", | |
| 427 | + "\n", | |
| 428 | + " model.compile(loss='mse',\n", | |
| 429 | + " optimizer=optimizer,\n", | |
| 430 | + " metrics=['mae', 'mse'])\n", | |
| 431 | + " return model" | |
| 432 | + ] | |
| 433 | + }, | |
| 434 | + { | |
| 435 | + "cell_type": "code", | |
| 436 | + "execution_count": 25, | |
| 437 | + "metadata": {}, | |
| 438 | + "outputs": [ | |
| 439 | + { | |
| 440 | + "name": "stderr", | |
| 441 | + "output_type": "stream", | |
| 442 | + "text": [ | |
| 443 | + "2022-02-17 16:36:25.433421: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA\n", | |
| 444 | + "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n" | |
| 445 | + ] | |
| 446 | + } | |
| 447 | + ], | |
| 448 | + "source": [ | |
| 449 | + "model = build_model()" | |
| 450 | + ] | |
| 451 | + }, | |
| 452 | + { | |
| 453 | + "cell_type": "code", | |
| 454 | + "execution_count": 26, | |
| 455 | + "metadata": {}, | |
| 456 | + "outputs": [ | |
| 457 | + { | |
| 458 | + "name": "stdout", | |
| 459 | + "output_type": "stream", | |
| 460 | + "text": [ | |
| 461 | + "Model: \"sequential\"\n", | |
| 462 | + "_________________________________________________________________\n", | |
| 463 | + " Layer (type) Output Shape Param # \n", | |
| 464 | + "=================================================================\n", | |
| 465 | + " dense (Dense) (None, 64) 640 \n", | |
| 466 | + " \n", | |
| 467 | + " dense_1 (Dense) (None, 64) 4160 \n", | |
| 468 | + " \n", | |
| 469 | + " dense_2 (Dense) (None, 1) 65 \n", | |
| 470 | + " \n", | |
| 471 | + "=================================================================\n", | |
| 472 | + "Total params: 4,865\n", | |
| 473 | + "Trainable params: 4,865\n", | |
| 474 | + "Non-trainable params: 0\n", | |
| 475 | + "_________________________________________________________________\n" | |
| 476 | + ] | |
| 477 | + } | |
| 478 | + ], | |
| 479 | + "source": [ | |
| 480 | + "model.summary()" | |
| 481 | + ] | |
| 482 | + }, | |
| 483 | + { | |
| 484 | + "cell_type": "code", | |
| 485 | + "execution_count": 27, | |
| 486 | + "metadata": {}, | |
| 487 | + "outputs": [ | |
| 488 | + { | |
| 489 | + "data": { | |
| 490 | + "text/plain": [ | |
| 491 | + "array([[-0.10623884],\n", | |
| 492 | + " [-0.0092403 ],\n", | |
| 493 | + " [-0.09777171],\n", | |
| 494 | + " [-0.0091932 ],\n", | |
| 495 | + " [-0.3904128 ],\n", | |
| 496 | + " [-0.06228563],\n", | |
| 497 | + " [-0.3566698 ],\n", | |
| 498 | + " [-0.5763702 ],\n", | |
| 499 | + " [-0.0279588 ],\n", | |
| 500 | + " [-0.40395904]], dtype=float32)" | |
| 501 | + ] | |
| 502 | + }, | |
| 503 | + "execution_count": 27, | |
| 504 | + "metadata": {}, | |
| 505 | + "output_type": "execute_result" | |
| 506 | + } | |
| 507 | + ], | |
| 508 | + "source": [ | |
| 509 | + "example_batch = normed_train_data[:10]\n", | |
| 510 | + "example_result = model.predict(example_batch)\n", | |
| 511 | + "example_result" | |
| 512 | + ] | |
| 513 | + }, | |
| 514 | + { | |
| 515 | + "cell_type": "code", | |
| 516 | + "execution_count": 28, | |
| 517 | + "metadata": {}, | |
| 518 | + "outputs": [ | |
| 519 | + { | |
| 520 | + "name": "stdout", | |
| 521 | + "output_type": "stream", | |
| 522 | + "text": [ | |
| 523 | + "\n", | |
| 524 | + "....................................................................................................\n", | |
| 525 | + "....................................................................................................\n", | |
| 526 | + "....................................................................................................\n", | |
| 527 | + "....................................................................................................\n", | |
| 528 | + "....................................................................................................\n", | |
| 529 | + "....................................................................................................\n", | |
| 530 | + "....................................................................................................\n", | |
| 531 | + "....................................................................................................\n", | |
| 532 | + "....................................................................................................\n", | |
| 533 | + "...................................................................................................." | |
| 534 | + ] | |
| 535 | + } | |
| 536 | + ], | |
| 537 | + "source": [ | |
| 538 | + "# ์ํฌํฌ๊ฐ ๋๋ ๋๋ง๋ค ์ (.)์ ์ถ๋ ฅํด ํ๋ จ ์งํ ๊ณผ์ ์ ํ์ํฉ๋๋ค\n", | |
| 539 | + "class PrintDot(keras.callbacks.Callback):\n", | |
| 540 | + " def on_epoch_end(self, epoch, logs):\n", | |
| 541 | + " if epoch % 100 == 0: print('')\n", | |
| 542 | + " print('.', end='')\n", | |
| 543 | + "\n", | |
| 544 | + "EPOCHS = 1000\n", | |
| 545 | + "\n", | |
| 546 | + "history = model.fit(\n", | |
| 547 | + " normed_train_data, train_labels,\n", | |
| 548 | + " epochs=EPOCHS, validation_split = 0.2, verbose=0,\n", | |
| 549 | + " callbacks=[PrintDot()])" | |
| 550 | + ] | |
| 551 | + }, | |
| 552 | + { | |
| 553 | + "cell_type": "code", | |
| 554 | + "execution_count": 29, | |
| 555 | + "metadata": {}, | |
| 556 | + "outputs": [ | |
| 557 | + { | |
| 558 | + "data": { | |
| 559 | + "text/html": [ | |
| 560 | + "<div>\n", | |
| 561 | + "<style scoped>\n", | |
| 562 | + " .dataframe tbody tr th:only-of-type {\n", | |
| 563 | + " vertical-align: middle;\n", | |
| 564 | + " }\n", | |
| 565 | + "\n", | |
| 566 | + " .dataframe tbody tr th {\n", | |
| 567 | + " vertical-align: top;\n", | |
| 568 | + " }\n", | |
| 569 | + "\n", | |
| 570 | + " .dataframe thead th {\n", | |
| 571 | + " text-align: right;\n", | |
| 572 | + " }\n", | |
| 573 | + "</style>\n", | |
| 574 | + "<table border=\"1\" class=\"dataframe\">\n", | |
| 575 | + " <thead>\n", | |
| 576 | + " <tr style=\"text-align: right;\">\n", | |
| 577 | + " <th></th>\n", | |
| 578 | + " <th>loss</th>\n", | |
| 579 | + " <th>mae</th>\n", | |
| 580 | + " <th>mse</th>\n", | |
| 581 | + " <th>val_loss</th>\n", | |
| 582 | + " <th>val_mae</th>\n", | |
| 583 | + " <th>val_mse</th>\n", | |
| 584 | + " <th>epoch</th>\n", | |
| 585 | + " </tr>\n", | |
| 586 | + " </thead>\n", | |
| 587 | + " <tbody>\n", | |
| 588 | + " <tr>\n", | |
| 589 | + " <th>995</th>\n", | |
| 590 | + " <td>2.500433</td>\n", | |
| 591 | + " <td>1.000197</td>\n", | |
| 592 | + " <td>2.500433</td>\n", | |
| 593 | + " <td>11.472550</td>\n", | |
| 594 | + " <td>2.508417</td>\n", | |
| 595 | + " <td>11.472550</td>\n", | |
| 596 | + " <td>995</td>\n", | |
| 597 | + " </tr>\n", | |
| 598 | + " <tr>\n", | |
| 599 | + " <th>996</th>\n", | |
| 600 | + " <td>2.405449</td>\n", | |
| 601 | + " <td>0.969858</td>\n", | |
| 602 | + " <td>2.405449</td>\n", | |
| 603 | + " <td>11.598408</td>\n", | |
| 604 | + " <td>2.482901</td>\n", | |
| 605 | + " <td>11.598408</td>\n", | |
| 606 | + " <td>996</td>\n", | |
| 607 | + " </tr>\n", | |
| 608 | + " <tr>\n", | |
| 609 | + " <th>997</th>\n", | |
| 610 | + " <td>2.468478</td>\n", | |
| 611 | + " <td>1.011330</td>\n", | |
| 612 | + " <td>2.468478</td>\n", | |
| 613 | + " <td>11.822487</td>\n", | |
| 614 | + " <td>2.496955</td>\n", | |
| 615 | + " <td>11.822487</td>\n", | |
| 616 | + " <td>997</td>\n", | |
| 617 | + " </tr>\n", | |
| 618 | + " <tr>\n", | |
| 619 | + " <th>998</th>\n", | |
| 620 | + " <td>2.478367</td>\n", | |
| 621 | + " <td>0.988899</td>\n", | |
| 622 | + " <td>2.478367</td>\n", | |
| 623 | + " <td>11.631916</td>\n", | |
| 624 | + " <td>2.503966</td>\n", | |
| 625 | + " <td>11.631916</td>\n", | |
| 626 | + " <td>998</td>\n", | |
| 627 | + " </tr>\n", | |
| 628 | + " <tr>\n", | |
| 629 | + " <th>999</th>\n", | |
| 630 | + " <td>2.387425</td>\n", | |
| 631 | + " <td>0.974386</td>\n", | |
| 632 | + " <td>2.387425</td>\n", | |
| 633 | + " <td>11.652516</td>\n", | |
| 634 | + " <td>2.505530</td>\n", | |
| 635 | + " <td>11.652516</td>\n", | |
| 636 | + " <td>999</td>\n", | |
| 637 | + " </tr>\n", | |
| 638 | + " </tbody>\n", | |
| 639 | + "</table>\n", | |
| 640 | + "</div>" | |
| 641 | + ], | |
| 642 | + "text/plain": [ | |
| 643 | + " loss mae mse val_loss val_mae val_mse epoch\n", | |
| 644 | + "995 2.500433 1.000197 2.500433 11.472550 2.508417 11.472550 995\n", | |
| 645 | + "996 2.405449 0.969858 2.405449 11.598408 2.482901 11.598408 996\n", | |
| 646 | + "997 2.468478 1.011330 2.468478 11.822487 2.496955 11.822487 997\n", | |
| 647 | + "998 2.478367 0.988899 2.478367 11.631916 2.503966 11.631916 998\n", | |
| 648 | + "999 2.387425 0.974386 2.387425 11.652516 2.505530 11.652516 999" | |
| 649 | + ] | |
| 650 | + }, | |
| 651 | + "execution_count": 29, | |
| 652 | + "metadata": {}, | |
| 653 | + "output_type": "execute_result" | |
| 654 | + } | |
| 655 | + ], | |
| 656 | + "source": [ | |
| 657 | + "hist = pd.DataFrame(history.history)\n", | |
| 658 | + "hist['epoch'] = history.epoch\n", | |
| 659 | + "hist.tail()" | |
| 660 | + ] | |
| 661 | + }, | |
| 662 | + { | |
| 663 | + "cell_type": "code", | |
| 664 | + "execution_count": 30, | |
| 665 | + "metadata": {}, | |
| 666 | + "outputs": [ | |
| 667 | + { | |
| 668 | + "data": { | |
| 669 | + "image/png": "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", | |
| 670 | + "text/plain": [ | |
| 671 | + "<Figure size 576x864 with 2 Axes>" | |
| 672 | + ] | |
| 673 | + }, | |
| 674 | + "metadata": { | |
| 675 | + "needs_background": "light" | |
| 676 | + }, | |
| 677 | + "output_type": "display_data" | |
| 678 | + } | |
| 679 | + ], | |
| 680 | + "source": [ | |
| 681 | + "import matplotlib.pyplot as plt\n", | |
| 682 | + "\n", | |
| 683 | + "def plot_history(history):\n", | |
| 684 | + " hist = pd.DataFrame(history.history)\n", | |
| 685 | + " hist['epoch'] = history.epoch\n", | |
| 686 | + "\n", | |
| 687 | + " plt.figure(figsize=(8,12))\n", | |
| 688 | + "\n", | |
| 689 | + " plt.subplot(2,1,1)\n", | |
| 690 | + " plt.xlabel('Epoch')\n", | |
| 691 | + " plt.ylabel('Mean Abs Error [MPG]')\n", | |
| 692 | + " plt.plot(hist['epoch'], hist['mae'],\n", | |
| 693 | + " label='Train Error')\n", | |
| 694 | + " plt.plot(hist['epoch'], hist['val_mae'],\n", | |
| 695 | + " label = 'Val Error')\n", | |
| 696 | + " plt.ylim([0,5])\n", | |
| 697 | + " plt.legend()\n", | |
| 698 | + "\n", | |
| 699 | + " plt.subplot(2,1,2)\n", | |
| 700 | + " plt.xlabel('Epoch')\n", | |
| 701 | + " plt.ylabel('Mean Square Error [$MPG^2$]')\n", | |
| 702 | + " plt.plot(hist['epoch'], hist['mse'],\n", | |
| 703 | + " label='Train Error')\n", | |
| 704 | + " plt.plot(hist['epoch'], hist['val_mse'],\n", | |
| 705 | + " label = 'Val Error')\n", | |
| 706 | + " plt.ylim([0,20])\n", | |
| 707 | + " plt.legend()\n", | |
| 708 | + " plt.show()\n", | |
| 709 | + "\n", | |
| 710 | + "plot_history(history)" | |
| 711 | + ] | |
| 712 | + }, | |
| 713 | + { | |
| 714 | + "cell_type": "code", | |
| 715 | + "execution_count": 31, | |
| 716 | + "metadata": {}, | |
| 717 | + "outputs": [ | |
| 718 | + { | |
| 719 | + "name": "stdout", | |
| 720 | + "output_type": "stream", | |
| 721 | + "text": [ | |
| 722 | + "\n", | |
| 723 | + "...................................................................." | |
| 724 | + ] | |
| 725 | + }, | |
| 726 | + { | |
| 727 | + "data": { | |
| 728 | + "image/png": "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", | |
| 729 | + "text/plain": [ | |
| 730 | + "<Figure size 576x864 with 2 Axes>" | |
| 731 | + ] | |
| 732 | + }, | |
| 733 | + "metadata": { | |
| 734 | + "needs_background": "light" | |
| 735 | + }, | |
| 736 | + "output_type": "display_data" | |
| 737 | + } | |
| 738 | + ], | |
| 739 | + "source": [ | |
| 740 | + "model = build_model()\n", | |
| 741 | + "\n", | |
| 742 | + "# patience ๋งค๊ฐ๋ณ์๋ ์ฑ๋ฅ ํฅ์์ ์ฒดํฌํ ์ํฌํฌ ํ์์ ๋๋ค\n", | |
| 743 | + "early_stop = keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)\n", | |
| 744 | + "\n", | |
| 745 | + "history = model.fit(normed_train_data, train_labels, epochs=EPOCHS,\n", | |
| 746 | + " validation_split = 0.2, verbose=0, callbacks=[early_stop, PrintDot()])\n", | |
| 747 | + "\n", | |
| 748 | + "plot_history(history)" | |
| 749 | + ] | |
| 750 | + }, | |
| 751 | + { | |
| 752 | + "cell_type": "code", | |
| 753 | + "execution_count": 32, | |
| 754 | + "metadata": {}, | |
| 755 | + "outputs": [ | |
| 756 | + { | |
| 757 | + "name": "stdout", | |
| 758 | + "output_type": "stream", | |
| 759 | + "text": [ | |
| 760 | + "3/3 - 0s - loss: 6.6843 - mae: 2.0619 - mse: 6.6843 - 17ms/epoch - 6ms/step\n", | |
| 761 | + "ํ ์คํธ ์ธํธ์ ํ๊ท ์ ๋ ์ค์ฐจ: 2.06 MPG\n" | |
| 762 | + ] | |
| 763 | + } | |
| 764 | + ], | |
| 765 | + "source": [ | |
| 766 | + "loss, mae, mse = model.evaluate(normed_test_data, test_labels, verbose=2)\n", | |
| 767 | + "\n", | |
| 768 | + "print(\"ํ ์คํธ ์ธํธ์ ํ๊ท ์ ๋ ์ค์ฐจ: {:5.2f} MPG\".format(mae))" | |
| 769 | + ] | |
| 770 | + }, | |
| 771 | + { | |
| 772 | + "cell_type": "code", | |
| 773 | + "execution_count": 33, | |
| 774 | + "metadata": {}, | |
| 775 | + "outputs": [ | |
| 776 | + { | |
| 777 | + "data": { | |
| 778 | + "image/png": "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", | |
| 779 | + "text/plain": [ | |
| 780 | + "<Figure size 432x288 with 1 Axes>" | |
| 781 | + ] | |
| 782 | + }, | |
| 783 | + "metadata": { | |
| 784 | + "needs_background": "light" | |
| 785 | + }, | |
| 786 | + "output_type": "display_data" | |
| 787 | + } | |
| 788 | + ], | |
| 789 | + "source": [ | |
| 790 | + "test_predictions = model.predict(normed_test_data).flatten()\n", | |
| 791 | + "\n", | |
| 792 | + "plt.scatter(test_labels, test_predictions)\n", | |
| 793 | + "plt.xlabel('True Values [MPG]')\n", | |
| 794 | + "plt.ylabel('Predictions [MPG]')\n", | |
| 795 | + "plt.axis('equal')\n", | |
| 796 | + "plt.axis('square')\n", | |
| 797 | + "plt.xlim([0,plt.xlim()[1]])\n", | |
| 798 | + "plt.ylim([0,plt.ylim()[1]])\n", | |
| 799 | + "_ = plt.plot([-100, 100], [-100, 100])" | |
| 800 | + ] | |
| 801 | + }, | |
| 802 | + { | |
| 803 | + "cell_type": "code", | |
| 804 | + "execution_count": 34, | |
| 805 | + "metadata": {}, | |
| 806 | + "outputs": [ | |
| 807 | + { | |
| 808 | + "data": { | |
| 809 | + "image/png": "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", | |
| 810 | + "text/plain": [ | |
| 811 | + "<Figure size 432x288 with 1 Axes>" | |
| 812 | + ] | |
| 813 | + }, | |
| 814 | + "metadata": { | |
| 815 | + "needs_background": "light" | |
| 816 | + }, | |
| 817 | + "output_type": "display_data" | |
| 818 | + } | |
| 819 | + ], | |
| 820 | + "source": [ | |
| 821 | + "error = test_predictions - test_labels\n", | |
| 822 | + "plt.hist(error, bins = 25)\n", | |
| 823 | + "plt.xlabel(\"Prediction Error [MPG]\")\n", | |
| 824 | + "_ = plt.ylabel(\"Count\")" | |
| 247 | 825 | ] |
| 248 | 826 | } |
| 249 | 827 | ], |