File size: 83,401 Bytes
57decc6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 |
{
"cells": [
{
"cell_type": "markdown",
"id": "c54f5831-58eb-4f9e-bb8a-2c2a6536a658",
"metadata": {},
"source": [
"# Denoising Diffusion Probabilistic Models with MedNIST Dataset Bundle \n",
"\n",
"This notebook discusses and uses the MONAI bundle it's included in for generating images from the MedNIST dataset using diffusion models. This is based off the 2d_ddpm_tutorial_ignite.ipynb notebook with a few changes.\n",
"\n",
"The bundle defines training and inference scripts whose use will be described here along with visualisations. The assumption with this notebook is that it's run within the bundle's `docs` directory and that the environment it runs in has `MONAI` installed. The command lines given are known to work in `bash` however may be problematic in Windows.\n",
"\n",
"Specifically, we train a diffusion model to generate X-Ray hands (drawn from the MedNIST dataset).\n",
"\n",
"First thing to do is import libraries and verify MONAI is present:"
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "6d32f8a4-2bfe-4cfb-9abd-033b0c6080e6",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"MONAI version: 1.4.0\n",
"Numpy version: 1.26.4\n",
"Pytorch version: 2.5.1+cu124\n",
"MONAI flags: HAS_EXT = False, USE_COMPILED = False, USE_META_DICT = False\n",
"MONAI rev id: 46a5272196a6c2590ca2589029eed8e4d56ff008\n",
"MONAI __file__: /data/PycharmProjects/monai-model-zoo/venv/lib/python3.10/site-packages/monai/__init__.py\n",
"\n",
"Optional dependencies:\n",
"Pytorch Ignite version: 0.5.1\n",
"ITK version: 5.4.0\n",
"Nibabel version: 5.3.2\n",
"scikit-image version: 0.25.0\n",
"scipy version: 1.15.1\n",
"Pillow version: 11.1.0\n",
"Tensorboard version: 2.18.0\n",
"gdown version: 5.2.0\n",
"TorchVision version: NOT INSTALLED or UNKNOWN VERSION.\n",
"tqdm version: 4.67.1\n",
"lmdb version: NOT INSTALLED or UNKNOWN VERSION.\n",
"psutil version: 6.1.1\n",
"pandas version: 2.2.3\n",
"einops version: 0.8.0\n",
"transformers version: 4.48.0\n",
"mlflow version: NOT INSTALLED or UNKNOWN VERSION.\n",
"pynrrd version: NOT INSTALLED or UNKNOWN VERSION.\n",
"clearml version: NOT INSTALLED or UNKNOWN VERSION.\n",
"\n",
"For details about installing the optional dependencies, please visit:\n",
" https://docs.monai.io/en/latest/installation.html#installing-the-recommended-dependencies\n",
"\n"
]
}
],
"source": [
"import os\n",
"import shutil\n",
"import tempfile\n",
"from pathlib import Path\n",
"\n",
"import torch\n",
"\n",
"import matplotlib.pyplot as plt\n",
"import monai\n",
"from monai.bundle import ConfigParser\n",
"\n",
"# path to the bundle directory, this assumes you're running the notebook in its directory\n",
"bundle_root = str(Path(\".\").absolute().parent)\n",
"\n",
"monai.config.print_config()"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "d6fc6592-cb51-4527-97ee-add5d1cdbeb4",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"/tmp/tmpdoc1dv_l\n"
]
}
],
"source": [
"directory = os.environ.get(\"MONAI_DATA_DIRECTORY\")\n",
"dataset_dir = tempfile.mkdtemp() if directory is None else directory\n",
"print(dataset_dir)"
]
},
{
"cell_type": "markdown",
"id": "5721b12a-8474-435b-aac2-c0ed054fa618",
"metadata": {},
"source": [
"### Training the diffusion model"
]
},
{
"cell_type": "markdown",
"id": "678d2e51-dc2d-4ad9-a4c0-14a6f900398b",
"metadata": {},
"source": [
"A bundle can be run on the command line using the Fire library or by parsing the configuration manually then getting parsed content objects. The following is the command to train the network for the default number of epochs. It will define values in the config files which need to be set for a particular run, such as the dataset directory created above, and setting the PYTHONPATH variable. The configuration for this bundle is split into 3 yaml files, one each for training (train.yaml) and inference (inference.yaml), and one to enable multi-GPU training (train_multigpu.yaml) which can be combined with the others to enable distributed training."
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "d52a4ae9-0d6d-4bc4-a5b5-f84470711f2d",
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"2025-01-28 10:45:10,893 - INFO - --- input summary of monai.bundle.scripts.run ---\n",
"2025-01-28 10:45:10,894 - INFO - > config_file: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/train.yaml'\n",
"2025-01-28 10:45:10,894 - INFO - > meta_file: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/metadata.json'\n",
"2025-01-28 10:45:10,894 - INFO - > logging_file: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/logging.conf'\n",
"2025-01-28 10:45:10,895 - INFO - > run_id: 'training'\n",
"2025-01-28 10:45:10,895 - INFO - > bundle_root: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm'\n",
"2025-01-28 10:45:10,895 - INFO - > dataset_dir: '/tmp/tmpdoc1dv_l'\n",
"2025-01-28 10:45:10,895 - INFO - > output_dir: './outputs'\n",
"2025-01-28 10:45:10,895 - INFO - ---\n",
"\n",
"\n",
"2025-01-28 10:45:10,896 - INFO - Setting logging properties based on config: /data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/logging.conf.\n",
"Detected deprecated name 'optional_packages_version' in configuration file, replacing with 'required_packages_version'.\n",
"2025-01-28 10:45:16,858 - INFO - Downloaded: /tmp/tmpdoc1dv_l/MedNIST.tar.gz\n",
"2025-01-28 10:45:16,933 - INFO - Verified 'MedNIST.tar.gz', md5: 0bc7306e7427e00ad1c5526a6677552d.\n",
"2025-01-28 10:45:16,934 - INFO - Writing into directory: /tmp/tmpdoc1dv_l.\n",
"2025-01-28 10:45:37,688 - INFO - Verified 'MedNIST.tar.gz', md5: 0bc7306e7427e00ad1c5526a6677552d.\n",
"2025-01-28 10:45:37,688 - INFO - File exists: /tmp/tmpdoc1dv_l/MedNIST.tar.gz, skipped downloading.\n",
"2025-01-28 10:45:37,688 - INFO - Non-empty folder exists in /tmp/tmpdoc1dv_l/MedNIST, skipped extracting.\n",
"`torch.cuda.amp.GradScaler(args...)` is deprecated. Please use `torch.amp.GradScaler('cuda', args...)` instead.\n",
"2025-01-28 10:45:39,645 - ignite.engine.engine.SupervisedTrainer - INFO - Engine run resuming from iteration 0, epoch 0 until 75 epochs\n",
"`torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.\n",
"2025-01-28 10:46:38,033 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.1607925146818161\n",
"2025-01-28 10:46:38,034 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[1] Complete. Time taken: 00:00:58.388\n",
"2025-01-28 10:47:29,911 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.016663629561662674\n",
"2025-01-28 10:47:29,911 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[2] Complete. Time taken: 00:00:51.877\n",
"2025-01-28 10:48:30,636 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.01578485034406185\n",
"2025-01-28 10:48:30,637 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[3] Complete. Time taken: 00:01:00.725\n",
"2025-01-28 10:49:23,157 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.013587715104222298\n",
"2025-01-28 10:49:23,157 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[4] Complete. Time taken: 00:00:52.521\n",
"2025-01-28 10:50:14,409 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.012479547411203384\n",
"2025-01-28 10:50:14,410 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 4 until 5 epochs\n",
"`torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.\n",
"2025-01-28 10:50:17,861 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.05754538252949715\n",
"2025-01-28 10:50:17,862 - INFO - Epoch[5] Metrics -- val_mean_abs_error: 0.0575 \n",
"2025-01-28 10:50:17,862 - INFO - Key metric: val_mean_abs_error best value: 0.05754538252949715 at epoch: 5\n",
"2025-01-28 10:50:17,862 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[5] Complete. Time taken: 00:00:03.373\n",
"2025-01-28 10:50:17,862 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.452\n",
"2025-01-28 10:50:17,933 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 5\n",
"2025-01-28 10:50:17,933 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[5] Complete. Time taken: 00:00:54.775\n",
"2025-01-28 10:51:11,088 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.012240087613463402\n",
"2025-01-28 10:51:11,088 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[6] Complete. Time taken: 00:00:53.155\n",
"2025-01-28 10:52:04,514 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[7] Complete. Time taken: 00:00:53.427\n",
"2025-01-28 10:52:57,317 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[8] Complete. Time taken: 00:00:52.803\n",
"2025-01-28 10:53:49,711 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.01170545443892479\n",
"2025-01-28 10:53:49,712 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[9] Complete. Time taken: 00:00:52.395\n",
"2025-01-28 10:54:51,839 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 9 until 10 epochs\n",
"2025-01-28 10:54:55,220 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.052069272845983505\n",
"2025-01-28 10:54:55,220 - INFO - Epoch[10] Metrics -- val_mean_abs_error: 0.0521 \n",
"2025-01-28 10:54:55,220 - INFO - Key metric: val_mean_abs_error best value: 0.052069272845983505 at epoch: 10\n",
"2025-01-28 10:54:55,220 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[10] Complete. Time taken: 00:00:03.381\n",
"2025-01-28 10:54:55,220 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.381\n",
"2025-01-28 10:54:55,292 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 10\n",
"2025-01-28 10:54:55,292 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[10] Complete. Time taken: 00:01:05.580\n",
"2025-01-28 10:55:49,249 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.011470048688352108\n",
"2025-01-28 10:55:49,249 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[11] Complete. Time taken: 00:00:53.957\n",
"2025-01-28 10:56:41,933 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.010766257531940937\n",
"2025-01-28 10:56:41,933 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[12] Complete. Time taken: 00:00:52.684\n",
"2025-01-28 10:57:32,681 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[13] Complete. Time taken: 00:00:50.748\n",
"2025-01-28 10:58:22,006 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.010334153659641743\n",
"2025-01-28 10:58:22,007 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[14] Complete. Time taken: 00:00:49.326\n",
"2025-01-28 10:59:12,718 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 14 until 15 epochs\n",
"2025-01-28 10:59:15,903 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.04713250324130058\n",
"2025-01-28 10:59:15,903 - INFO - Epoch[15] Metrics -- val_mean_abs_error: 0.0471 \n",
"2025-01-28 10:59:15,903 - INFO - Key metric: val_mean_abs_error best value: 0.04713250324130058 at epoch: 15\n",
"2025-01-28 10:59:15,903 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[15] Complete. Time taken: 00:00:03.184\n",
"2025-01-28 10:59:15,903 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.185\n",
"2025-01-28 10:59:15,972 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 15\n",
"2025-01-28 10:59:15,972 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[15] Complete. Time taken: 00:00:53.966\n",
"2025-01-28 11:00:05,379 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.010036583058536053\n",
"2025-01-28 11:00:05,379 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[16] Complete. Time taken: 00:00:49.407\n",
"2025-01-28 11:00:54,954 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[17] Complete. Time taken: 00:00:49.575\n",
"2025-01-28 11:01:44,481 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[18] Complete. Time taken: 00:00:49.527\n",
"2025-01-28 11:02:34,070 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.010024736635386944\n",
"2025-01-28 11:02:34,070 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[19] Complete. Time taken: 00:00:49.590\n",
"2025-01-28 11:03:24,598 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 19 until 20 epochs\n",
"2025-01-28 11:03:28,013 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.04626006633043289\n",
"2025-01-28 11:03:28,013 - INFO - Epoch[20] Metrics -- val_mean_abs_error: 0.0463 \n",
"2025-01-28 11:03:28,013 - INFO - Key metric: val_mean_abs_error best value: 0.04626006633043289 at epoch: 20\n",
"2025-01-28 11:03:28,013 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[20] Complete. Time taken: 00:00:03.414\n",
"2025-01-28 11:03:28,014 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.415\n",
"2025-01-28 11:03:28,373 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 20\n",
"2025-01-28 11:03:28,373 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[20] Complete. Time taken: 00:00:54.303\n",
"2025-01-28 11:04:21,035 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[21] Complete. Time taken: 00:00:52.662\n",
"2025-01-28 11:05:11,684 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.010010532103478909\n",
"2025-01-28 11:05:11,684 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[22] Complete. Time taken: 00:00:50.648\n",
"2025-01-28 11:06:00,999 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.0098584508523345\n",
"2025-01-28 11:06:01,000 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[23] Complete. Time taken: 00:00:49.316\n",
"2025-01-28 11:06:54,731 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[24] Complete. Time taken: 00:00:53.732\n",
"2025-01-28 11:07:49,033 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 24 until 25 epochs\n",
"2025-01-28 11:07:52,450 - INFO - Epoch[25] Metrics -- val_mean_abs_error: 0.0496 \n",
"2025-01-28 11:07:52,450 - INFO - Key metric: val_mean_abs_error best value: 0.04626006633043289 at epoch: 20\n",
"2025-01-28 11:07:52,450 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[25] Complete. Time taken: 00:00:03.416\n",
"2025-01-28 11:07:52,450 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.417\n",
"2025-01-28 11:07:52,809 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 25\n",
"2025-01-28 11:07:52,809 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[25] Complete. Time taken: 00:00:58.078\n",
"2025-01-28 11:08:47,931 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.00983799621462822\n",
"2025-01-28 11:08:47,931 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[26] Complete. Time taken: 00:00:55.122\n",
"2025-01-28 11:09:40,163 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.009661602787673473\n",
"2025-01-28 11:09:40,163 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[27] Complete. Time taken: 00:00:52.232\n",
"2025-01-28 11:10:32,018 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[28] Complete. Time taken: 00:00:51.855\n",
"2025-01-28 11:11:26,697 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[29] Complete. Time taken: 00:00:54.679\n",
"2025-01-28 11:12:19,319 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 29 until 30 epochs\n",
"2025-01-28 11:12:22,592 - INFO - Epoch[30] Metrics -- val_mean_abs_error: 0.0470 \n",
"2025-01-28 11:12:22,592 - INFO - Key metric: val_mean_abs_error best value: 0.04626006633043289 at epoch: 20\n",
"2025-01-28 11:12:22,592 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[30] Complete. Time taken: 00:00:03.272\n",
"2025-01-28 11:12:22,592 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.273\n",
"2025-01-28 11:12:22,649 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 30\n",
"2025-01-28 11:12:22,649 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[30] Complete. Time taken: 00:00:55.952\n",
"2025-01-28 11:13:16,180 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.00965067371726036\n",
"2025-01-28 11:13:16,180 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[31] Complete. Time taken: 00:00:53.531\n",
"2025-01-28 11:14:10,456 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.009442757815122604\n",
"2025-01-28 11:14:10,456 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[32] Complete. Time taken: 00:00:54.275\n",
"2025-01-28 11:15:03,456 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.008967726491391659\n",
"2025-01-28 11:15:03,456 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[33] Complete. Time taken: 00:00:53.000\n",
"2025-01-28 11:15:55,835 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[34] Complete. Time taken: 00:00:52.379\n",
"2025-01-28 11:16:47,540 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 34 until 35 epochs\n",
"2025-01-28 11:16:50,823 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.04337985813617706\n",
"2025-01-28 11:16:50,823 - INFO - Epoch[35] Metrics -- val_mean_abs_error: 0.0434 \n",
"2025-01-28 11:16:50,823 - INFO - Key metric: val_mean_abs_error best value: 0.04337985813617706 at epoch: 35\n",
"2025-01-28 11:16:50,823 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[35] Complete. Time taken: 00:00:03.282\n",
"2025-01-28 11:16:50,823 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.283\n",
"2025-01-28 11:16:50,894 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 35\n",
"2025-01-28 11:16:50,894 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[35] Complete. Time taken: 00:00:55.060\n",
"2025-01-28 11:17:42,195 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[36] Complete. Time taken: 00:00:51.301\n",
"2025-01-28 11:18:34,921 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[37] Complete. Time taken: 00:00:52.726\n",
"2025-01-28 11:19:27,769 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[38] Complete. Time taken: 00:00:52.848\n",
"2025-01-28 11:20:20,623 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[39] Complete. Time taken: 00:00:52.854\n",
"2025-01-28 11:21:11,887 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 39 until 40 epochs\n",
"2025-01-28 11:21:14,992 - INFO - Epoch[40] Metrics -- val_mean_abs_error: 0.0438 \n",
"2025-01-28 11:21:14,992 - INFO - Key metric: val_mean_abs_error best value: 0.04337985813617706 at epoch: 35\n",
"2025-01-28 11:21:14,992 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[40] Complete. Time taken: 00:00:03.104\n",
"2025-01-28 11:21:14,992 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.105\n",
"2025-01-28 11:21:15,062 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 40\n",
"2025-01-28 11:21:15,062 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[40] Complete. Time taken: 00:00:54.439\n",
"2025-01-28 11:22:06,367 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[41] Complete. Time taken: 00:00:51.305\n",
"2025-01-28 11:22:57,686 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[42] Complete. Time taken: 00:00:51.319\n",
"2025-01-28 11:23:49,550 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[43] Complete. Time taken: 00:00:51.864\n",
"2025-01-28 11:24:40,625 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[44] Complete. Time taken: 00:00:51.074\n",
"2025-01-28 11:25:31,099 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 44 until 45 epochs\n",
"2025-01-28 11:25:34,234 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.04306837171316147\n",
"2025-01-28 11:25:34,234 - INFO - Epoch[45] Metrics -- val_mean_abs_error: 0.0431 \n",
"2025-01-28 11:25:34,234 - INFO - Key metric: val_mean_abs_error best value: 0.04306837171316147 at epoch: 45\n",
"2025-01-28 11:25:34,234 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[45] Complete. Time taken: 00:00:03.135\n",
"2025-01-28 11:25:34,235 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.135\n",
"2025-01-28 11:25:34,584 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 45\n",
"2025-01-28 11:25:34,585 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[45] Complete. Time taken: 00:00:53.960\n",
"2025-01-28 11:26:24,509 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[46] Complete. Time taken: 00:00:49.924\n",
"2025-01-28 11:27:13,455 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[47] Complete. Time taken: 00:00:48.945\n",
"2025-01-28 11:28:03,014 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[48] Complete. Time taken: 00:00:49.559\n",
"2025-01-28 11:28:52,035 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[49] Complete. Time taken: 00:00:49.021\n",
"2025-01-28 11:29:41,226 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 49 until 50 epochs\n",
"2025-01-28 11:29:44,269 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.0430283285677433\n",
"2025-01-28 11:29:44,269 - INFO - Epoch[50] Metrics -- val_mean_abs_error: 0.0430 \n",
"2025-01-28 11:29:44,269 - INFO - Key metric: val_mean_abs_error best value: 0.0430283285677433 at epoch: 50\n",
"2025-01-28 11:29:44,269 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[50] Complete. Time taken: 00:00:03.042\n",
"2025-01-28 11:29:44,269 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.043\n",
"2025-01-28 11:29:44,628 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 50\n",
"2025-01-28 11:29:44,628 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[50] Complete. Time taken: 00:00:52.594\n",
"2025-01-28 11:30:34,707 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[51] Complete. Time taken: 00:00:50.079\n",
"2025-01-28 11:31:25,426 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.008929001167416573\n",
"2025-01-28 11:31:25,426 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[52] Complete. Time taken: 00:00:50.719\n",
"2025-01-28 11:32:17,793 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.008782487362623215\n",
"2025-01-28 11:32:17,794 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[53] Complete. Time taken: 00:00:52.368\n",
"2025-01-28 11:33:09,429 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.008487475104629993\n",
"2025-01-28 11:33:09,429 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[54] Complete. Time taken: 00:00:51.636\n",
"2025-01-28 11:34:00,926 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 54 until 55 epochs\n",
"2025-01-28 11:34:04,209 - INFO - Epoch[55] Metrics -- val_mean_abs_error: 0.0439 \n",
"2025-01-28 11:34:04,209 - INFO - Key metric: val_mean_abs_error best value: 0.0430283285677433 at epoch: 50\n",
"2025-01-28 11:34:04,209 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[55] Complete. Time taken: 00:00:03.283\n",
"2025-01-28 11:34:04,209 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.283\n",
"2025-01-28 11:34:04,277 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 55\n",
"2025-01-28 11:34:04,277 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[55] Complete. Time taken: 00:00:54.848\n",
"2025-01-28 11:34:55,848 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[56] Complete. Time taken: 00:00:51.571\n",
"2025-01-28 11:35:47,270 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[57] Complete. Time taken: 00:00:51.422\n",
"2025-01-28 11:36:37,430 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[58] Complete. Time taken: 00:00:50.160\n",
"2025-01-28 11:37:27,691 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[59] Complete. Time taken: 00:00:50.261\n",
"2025-01-28 11:38:19,228 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 59 until 60 epochs\n",
"2025-01-28 11:38:22,427 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.041947875171899796\n",
"2025-01-28 11:38:22,427 - INFO - Epoch[60] Metrics -- val_mean_abs_error: 0.0419 \n",
"2025-01-28 11:38:22,427 - INFO - Key metric: val_mean_abs_error best value: 0.041947875171899796 at epoch: 60\n",
"2025-01-28 11:38:22,427 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[60] Complete. Time taken: 00:00:03.198\n",
"2025-01-28 11:38:22,427 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.198\n",
"2025-01-28 11:38:22,785 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 60\n",
"2025-01-28 11:38:22,785 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[60] Complete. Time taken: 00:00:55.094\n",
"2025-01-28 11:39:14,371 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[61] Complete. Time taken: 00:00:51.586\n",
"2025-01-28 11:40:06,623 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[62] Complete. Time taken: 00:00:52.252\n",
"2025-01-28 11:40:58,129 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[63] Complete. Time taken: 00:00:51.506\n",
"2025-01-28 11:41:47,897 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[64] Complete. Time taken: 00:00:49.768\n",
"2025-01-28 11:42:39,507 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 64 until 65 epochs\n",
"2025-01-28 11:42:42,800 - ignite.engine.engine.SupervisedEvaluator - INFO - Got new best metric of val_mean_abs_error: 0.03936021775007248\n",
"2025-01-28 11:42:42,800 - INFO - Epoch[65] Metrics -- val_mean_abs_error: 0.0394 \n",
"2025-01-28 11:42:42,800 - INFO - Key metric: val_mean_abs_error best value: 0.03936021775007248 at epoch: 65\n",
"2025-01-28 11:42:42,800 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[65] Complete. Time taken: 00:00:03.292\n",
"2025-01-28 11:42:42,800 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.293\n",
"2025-01-28 11:42:42,907 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 65\n",
"2025-01-28 11:42:42,907 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[65] Complete. Time taken: 00:00:55.010\n",
"2025-01-28 11:43:34,027 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[66] Complete. Time taken: 00:00:51.120\n",
"2025-01-28 11:44:23,357 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[67] Complete. Time taken: 00:00:49.330\n",
"2025-01-28 11:45:15,222 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[68] Complete. Time taken: 00:00:51.865\n",
"2025-01-28 11:46:05,380 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[69] Complete. Time taken: 00:00:50.158\n",
"2025-01-28 11:46:54,867 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 69 until 70 epochs\n",
"2025-01-28 11:46:58,033 - INFO - Epoch[70] Metrics -- val_mean_abs_error: 0.0419 \n",
"2025-01-28 11:46:58,033 - INFO - Key metric: val_mean_abs_error best value: 0.03936021775007248 at epoch: 65\n",
"2025-01-28 11:46:58,033 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[70] Complete. Time taken: 00:00:03.165\n",
"2025-01-28 11:46:58,033 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.166\n",
"2025-01-28 11:46:58,103 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 70\n",
"2025-01-28 11:46:58,104 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[70] Complete. Time taken: 00:00:52.723\n",
"2025-01-28 11:47:48,003 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[71] Complete. Time taken: 00:00:49.899\n",
"2025-01-28 11:48:39,114 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[72] Complete. Time taken: 00:00:51.111\n",
"2025-01-28 11:49:28,678 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[73] Complete. Time taken: 00:00:49.564\n",
"2025-01-28 11:50:20,176 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[74] Complete. Time taken: 00:00:51.498\n",
"2025-01-28 11:51:10,599 - ignite.engine.engine.SupervisedTrainer - INFO - Got new best metric of train_acc: 0.008344327099621296\n",
"2025-01-28 11:51:10,599 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run resuming from iteration 0, epoch 74 until 75 epochs\n",
"2025-01-28 11:51:13,917 - INFO - Epoch[75] Metrics -- val_mean_abs_error: 0.0419 \n",
"2025-01-28 11:51:13,918 - INFO - Key metric: val_mean_abs_error best value: 0.03936021775007248 at epoch: 65\n",
"2025-01-28 11:51:13,918 - ignite.engine.engine.SupervisedEvaluator - INFO - Epoch[75] Complete. Time taken: 00:00:03.318\n",
"2025-01-28 11:51:13,918 - ignite.engine.engine.SupervisedEvaluator - INFO - Engine run complete. Time taken: 00:00:03.319\n",
"2025-01-28 11:51:13,990 - ignite.engine.engine.SupervisedTrainer - INFO - Saved checkpoint at epoch: 75\n",
"2025-01-28 11:51:13,990 - ignite.engine.engine.SupervisedTrainer - INFO - Epoch[75] Complete. Time taken: 00:00:53.813\n",
"2025-01-28 11:51:14,103 - ignite.engine.engine.SupervisedTrainer - INFO - Train completed, saved final checkpoint: outputs/model_final_iteration=75000.pt\n",
"2025-01-28 11:51:14,103 - ignite.engine.engine.SupervisedTrainer - INFO - Engine run complete. Time taken: 01:05:34.457\n"
]
}
],
"source": [
"# multiple config files need to be specified this way with '' quotes, variable used in command line must be in \"\" quotes\n",
"configs=f\"{bundle_root}/configs/train.yaml\"\n",
"!PYTHONPATH={bundle_root} python -m monai.bundle run training \\\n",
" --meta_file {bundle_root}/configs/metadata.json \\\n",
" --config_file \"{configs}\" \\\n",
" --logging_file {bundle_root}/configs/logging.conf \\\n",
" --bundle_root {bundle_root} \\\n",
" --dataset_dir {dataset_dir} \\\n",
" --output_dir './outputs'"
]
},
{
"cell_type": "markdown",
"id": "f872fccf-12af-43ef-bdb5-f49bc74119fa",
"metadata": {},
"source": [
"### Test the diffusion model"
]
},
{
"cell_type": "markdown",
"id": "5030732c-deb5-448a-b575-385bda0fa308",
"metadata": {},
"source": [
"The test inference script can then be invoked as such to produce an output tensor saved to the given file with a randomly generated image. The `ckpt_path` value should point to the final checkpoint file created during the above training run which will be in a subdirectory of `./results`. The training script's default behaviour is to create a new timestamped subdirectory in `./results` for every new run, this can be explicitly set by providing a `output_dir` value on the command line."
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "40e6a3e9-3984-44b0-ba9a-5b8d58c7ea2d",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"2025-01-28 11:52:00,179 - INFO - --- input summary of monai.bundle.scripts.run ---\n",
"2025-01-28 11:52:00,179 - INFO - > config_file: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/inference.yaml'\n",
"2025-01-28 11:52:00,179 - INFO - > meta_file: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/metadata.json'\n",
"2025-01-28 11:52:00,179 - INFO - > run_id: 'testing'\n",
"2025-01-28 11:52:00,179 - INFO - > ckpt_path: './outputs/model_final_iteration=75000.pt'\n",
"2025-01-28 11:52:00,179 - INFO - > bundle_root: '/data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm'\n",
"2025-01-28 11:52:00,179 - INFO - > out_file: 'test.pt'\n",
"2025-01-28 11:52:00,179 - INFO - ---\n",
"\n",
"\n",
"2025-01-28 11:52:00,179 - INFO - Setting logging properties based on config: /data/PycharmProjects/monai-model-zoo/model-zoo/models/mednist_ddpm/configs/logging.conf.\n",
"Detected deprecated name 'optional_packages_version' in configuration file, replacing with 'required_packages_version'.\n",
"100%|██████████████████████████████████████| 1000/1000 [00:08<00:00, 113.53it/s]\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n"
]
},
{
"data": {
"text/plain": [
"<matplotlib.image.AxesImage at 0x7f5db4c46050>"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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",
"text/plain": [
"<Figure size 640x480 with 1 Axes>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"configs=f\"{bundle_root}/configs/inference.yaml\"\n",
"\n",
"!PYTHONPATH={bundle_root} python -m monai.bundle run testing \\\n",
" --meta_file {bundle_root}/configs/metadata.json \\\n",
" --config_file \"{configs}\" \\\n",
" --ckpt_path \"./outputs/model_final_iteration=75000.pt\" \\\n",
" --bundle_root {bundle_root} \\\n",
" --out_file test.pt\n",
"\n",
"test = torch.load(\"test.pt\", map_location=\"cpu\")\n",
"\n",
"plt.imshow(test[0, 0], vmin=0.4, vmax=1.0, cmap=\"gray\")"
]
},
{
"cell_type": "markdown",
"id": "f581c36e-4033-4005-8969-76205470588e",
"metadata": {},
"source": [
"The same can be done by creating the parser object, filling in its configuration, then resolving the Python objects from the constructed bundle data:"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "cf8438b3-4c7d-48c4-bb41-ed7def73753f",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1000/1000 [00:08<00:00, 122.14it/s]\n"
]
},
{
"data": {
"text/plain": [
"<matplotlib.image.AxesImage at 0x7f5dac07b220>"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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",
"text/plain": [
"<Figure size 640x480 with 1 Axes>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"import sys\n",
"\n",
"sys.path.append(bundle_root) # make sure we load the script files we need\n",
"\n",
"# configure the parser from the bundle's information\n",
"cp = ConfigParser()\n",
"cp.read_meta(f\"{bundle_root}/configs/metadata.json\")\n",
"cp.read_config([f\"{bundle_root}/configs/inference.yaml\"])\n",
"cp[\"bundle_root\"] = bundle_root\n",
"cp[\"ckpt_path\"] = f\"./outputs/model_final_iteration=75000.pt\"\n",
"\n",
"cp.get_parsed_content(\"load_state\") # load the saved state from the checkpoint just be resolving this value\n",
"\n",
"device = cp.get_parsed_content(\"device\") # device used by the bundle\n",
"sample = cp.get_parsed_content(\"sample\") # test sampling function\n",
"\n",
"image_dim = cp[\"image_dim\"] # get the stored dimension value, no need to resolve anything\n",
"\n",
"noise = torch.rand(1, 1, image_dim, image_dim).to(device) # or cp.get_parsed_content(\"noise\")\n",
"\n",
"test = sample(noise)\n",
"\n",
"plt.imshow(test[0, 0].cpu(), vmin=0.75, vmax=1, cmap=\"gray\") # Toggle vmin for more contrast"
]
},
{
"cell_type": "markdown",
"id": "2feab4e5-2745-4d35-9eec-a2bb8340cf51",
"metadata": {},
"source": [
"Multi-GPU can be enabled by including the `train_multigpu.yaml` configuration file:"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "173cda1c-ac90-410f-b34d-b6cbb0044c7a",
"metadata": {},
"outputs": [],
"source": [
"configs=f\"'{bundle_root}/configs/train.yaml', '{bundle_root}/configs/train_multigpu.yaml'\"\n",
"\n",
"!PYTHONPATH={bundle_root} torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run training \\\n",
" --meta_file {bundle_root}/configs/metadata.json \\\n",
" --config_file \"{configs}\" \\\n",
" --logging_file {bundle_root}/configs/logging.conf \\\n",
" --bundle_root {bundle_root} \\\n",
" --dataset_dir {dataset_dir}"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "cb719023-8250-43c4-ab10-911829332498",
"metadata": {},
"outputs": [],
"source": [
"if directory is None:\n",
" shutil.rmtree(dataset_dir)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"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": "3.10.16"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|