File size: 41,056 Bytes
17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 f237e67 17643e1 e52e688 f237e67 e52e688 f237e67 e52e688 17643e1 810e3d8 17643e1 f237e67 17643e1 810e3d8 17643e1 810e3d8 17643e1 f237e67 17643e1 810e3d8 17643e1 810e3d8 17643e1 f237e67 17643e1 f237e67 17643e1 810e3d8 17643e1 f237e67 17643e1 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 |
{
"cells": [
{
"cell_type": "markdown",
"id": "63b957e0-d83b-48a6-8ae0-a276b983e181",
"metadata": {},
"source": [
"### Optional: install the necessary packages"
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "5569a330-ea6b-4402-9fd5-e7a0ce981bc9",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Requirement already satisfied: huggingface_hub in /workspace/.miniconda3/lib/python3.12/site-packages (0.26.2)\n",
"Requirement already satisfied: filelock in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (3.16.1)\n",
"Requirement already satisfied: fsspec>=2023.5.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (2024.9.0)\n",
"Requirement already satisfied: packaging>=20.9 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (24.1)\n",
"Requirement already satisfied: pyyaml>=5.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (6.0.2)\n",
"Requirement already satisfied: requests in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (2.32.3)\n",
"Requirement already satisfied: tqdm>=4.42.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (4.66.5)\n",
"Requirement already satisfied: typing-extensions>=3.7.4.3 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface_hub) (4.11.0)\n",
"Requirement already satisfied: charset-normalizer<4,>=2 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests->huggingface_hub) (3.3.2)\n",
"Requirement already satisfied: idna<4,>=2.5 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests->huggingface_hub) (3.7)\n",
"Requirement already satisfied: urllib3<3,>=1.21.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests->huggingface_hub) (2.2.3)\n",
"Requirement already satisfied: certifi>=2017.4.17 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests->huggingface_hub) (2024.8.30)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Requirement already satisfied: datasets in /workspace/.miniconda3/lib/python3.12/site-packages (3.1.0)\n",
"Requirement already satisfied: filelock in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (3.16.1)\n",
"Requirement already satisfied: numpy>=1.17 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (2.1.3)\n",
"Requirement already satisfied: pyarrow>=15.0.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (18.0.0)\n",
"Requirement already satisfied: dill<0.3.9,>=0.3.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (0.3.8)\n",
"Requirement already satisfied: pandas in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (2.2.3)\n",
"Requirement already satisfied: requests>=2.32.2 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (2.32.3)\n",
"Requirement already satisfied: tqdm>=4.66.3 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (4.66.5)\n",
"Requirement already satisfied: xxhash in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (3.5.0)\n",
"Requirement already satisfied: multiprocess<0.70.17 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (0.70.16)\n",
"Requirement already satisfied: fsspec<=2024.9.0,>=2023.1.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from fsspec[http]<=2024.9.0,>=2023.1.0->datasets) (2024.9.0)\n",
"Requirement already satisfied: aiohttp in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (3.11.2)\n",
"Requirement already satisfied: huggingface-hub>=0.23.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (0.26.2)\n",
"Requirement already satisfied: packaging in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (24.1)\n",
"Requirement already satisfied: pyyaml>=5.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from datasets) (6.0.2)\n",
"Requirement already satisfied: aiohappyeyeballs>=2.3.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (2.4.3)\n",
"Requirement already satisfied: aiosignal>=1.1.2 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (1.3.1)\n",
"Requirement already satisfied: attrs>=17.3.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (24.2.0)\n",
"Requirement already satisfied: frozenlist>=1.1.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (1.5.0)\n",
"Requirement already satisfied: multidict<7.0,>=4.5 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (6.1.0)\n",
"Requirement already satisfied: propcache>=0.2.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (0.2.0)\n",
"Requirement already satisfied: yarl<2.0,>=1.17.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from aiohttp->datasets) (1.17.1)\n",
"Requirement already satisfied: typing-extensions>=3.7.4.3 in /workspace/.miniconda3/lib/python3.12/site-packages (from huggingface-hub>=0.23.0->datasets) (4.11.0)\n",
"Requirement already satisfied: charset-normalizer<4,>=2 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests>=2.32.2->datasets) (3.3.2)\n",
"Requirement already satisfied: idna<4,>=2.5 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests>=2.32.2->datasets) (3.7)\n",
"Requirement already satisfied: urllib3<3,>=1.21.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests>=2.32.2->datasets) (2.2.3)\n",
"Requirement already satisfied: certifi>=2017.4.17 in /workspace/.miniconda3/lib/python3.12/site-packages (from requests>=2.32.2->datasets) (2024.8.30)\n",
"Requirement already satisfied: python-dateutil>=2.8.2 in /workspace/.miniconda3/lib/python3.12/site-packages (from pandas->datasets) (2.9.0.post0)\n",
"Requirement already satisfied: pytz>=2020.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from pandas->datasets) (2024.1)\n",
"Requirement already satisfied: tzdata>=2022.7 in /workspace/.miniconda3/lib/python3.12/site-packages (from pandas->datasets) (2024.2)\n",
"Requirement already satisfied: six>=1.5 in /workspace/.miniconda3/lib/python3.12/site-packages (from python-dateutil>=2.8.2->pandas->datasets) (1.16.0)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Requirement already satisfied: bitsandbytes in /workspace/.miniconda3/lib/python3.12/site-packages (0.44.2.dev0)\n",
"Requirement already satisfied: torch in /workspace/.miniconda3/lib/python3.12/site-packages (from bitsandbytes) (2.5.1)\n",
"Requirement already satisfied: numpy in /workspace/.miniconda3/lib/python3.12/site-packages (from bitsandbytes) (2.1.3)\n",
"Requirement already satisfied: filelock in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (3.16.1)\n",
"Requirement already satisfied: typing-extensions>=4.8.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (4.11.0)\n",
"Requirement already satisfied: networkx in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (3.4.2)\n",
"Requirement already satisfied: jinja2 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (3.1.4)\n",
"Requirement already satisfied: fsspec in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (2024.9.0)\n",
"Requirement already satisfied: nvidia-cuda-nvrtc-cu12==12.4.127 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.127)\n",
"Requirement already satisfied: nvidia-cuda-runtime-cu12==12.4.127 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.127)\n",
"Requirement already satisfied: nvidia-cuda-cupti-cu12==12.4.127 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.127)\n",
"Requirement already satisfied: nvidia-cudnn-cu12==9.1.0.70 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (9.1.0.70)\n",
"Requirement already satisfied: nvidia-cublas-cu12==12.4.5.8 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.5.8)\n",
"Requirement already satisfied: nvidia-cufft-cu12==11.2.1.3 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (11.2.1.3)\n",
"Requirement already satisfied: nvidia-curand-cu12==10.3.5.147 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (10.3.5.147)\n",
"Requirement already satisfied: nvidia-cusolver-cu12==11.6.1.9 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (11.6.1.9)\n",
"Requirement already satisfied: nvidia-cusparse-cu12==12.3.1.170 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.3.1.170)\n",
"Requirement already satisfied: nvidia-nccl-cu12==2.21.5 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (2.21.5)\n",
"Requirement already satisfied: nvidia-nvtx-cu12==12.4.127 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.127)\n",
"Requirement already satisfied: nvidia-nvjitlink-cu12==12.4.127 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (12.4.127)\n",
"Requirement already satisfied: triton==3.1.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (3.1.0)\n",
"Requirement already satisfied: setuptools in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (75.1.0)\n",
"Requirement already satisfied: sympy==1.13.1 in /workspace/.miniconda3/lib/python3.12/site-packages (from torch->bitsandbytes) (1.13.1)\n",
"Requirement already satisfied: mpmath<1.4,>=1.1.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from sympy==1.13.1->torch->bitsandbytes) (1.3.0)\n",
"Requirement already satisfied: MarkupSafe>=2.0 in /workspace/.miniconda3/lib/python3.12/site-packages (from jinja2->torch->bitsandbytes) (2.1.3)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Note: you may need to restart the kernel to use updated packages.\n"
]
}
],
"source": [
"!git config --global credential.helper store\n",
"%pip install huggingface_hub\n",
"%pip install -U datasets\n",
"%pip install -U bitsandbytes\n",
"%pip install -q git+https://github.com/huggingface/transformers.git\n",
"%pip install -q accelerate datasets peft torchvision torchaudio"
]
},
{
"cell_type": "markdown",
"id": "ca7f5161-0104-45c4-83c9-1dd0dad15e29",
"metadata": {},
"source": [
"## Login on Hugging Face"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "14c445aa-c8bc-43b9-8d18-46d79055e1f0",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Hugging Face token found in environment variable\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n"
]
}
],
"source": [
"from huggingface_hub import login\n",
"import os\n",
"\n",
"HF_TOKEN = \"hf_C…………\"\n",
"\n",
"if os.environ.get('HF_TOKEN') is not None:\n",
" HF_TOKEN = os.environ.get('HF_TOKEN')\n",
" print(f\"Hugging Face token found in environment variable\")\n",
"try:\n",
" import google.colab\n",
" from google.colab import userdata\n",
" if (userdata.get('HF_TOKEN') is not None) and (HF_TOKEN == \"\"):\n",
" HF_TOKEN = userdata.get('HF_TOKEN')\n",
" else:\n",
" raise ValueError(\"Please set your Hugging Face token in the user data panel, or pass it as an environment variable\")\n",
"except ModuleNotFoundError:\n",
" if HF_TOKEN is None:\n",
" raise ValueError(\"Please set your Hugging Face token in the user data panel, or pass it as an environment variable\")\n",
"\n",
"login(\n",
" token=HF_TOKEN,\n",
" add_to_git_credential=True\n",
")"
]
},
{
"cell_type": "markdown",
"id": "c33a0325-f056-43c4-a500-6ad1b5632ee5",
"metadata": {},
"source": [
"### Set the environment variables"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "bfd9dbb4-59c4-4d97-9727-5c2d9e60724e",
"metadata": {},
"outputs": [],
"source": [
"#source_model_id = \"HuggingFaceM4/Idefics3-8B-Llama3\"\n",
"source_model_id = \"meta-llama/Llama-3.2-11B-Vision-Instruct\"\n",
"detination_model_id = \"eltorio/IDEFICS3_medical_instruct\"\n",
"dataset_id = \"ruslanmv/ai-medical-dataset\"\n",
"prompt= \"You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor\"\n",
"output_dir = \"IDEFICS3_medical_instruct\""
]
},
{
"cell_type": "markdown",
"id": "da89fd1d-44dd-4d4e-803f-ebf58b23165f",
"metadata": {},
"source": [
"### Optionally clone the model repository"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "f6510468-c903-4bd2-80fd-2511b7fb2f72",
"metadata": {},
"outputs": [],
"source": [
"# clone Hugging Face model repository\n",
"# !git clone https://huggingface.co/$destination_model_id $output_dir"
]
},
{
"cell_type": "markdown",
"id": "2bfc04cd-3867-445f-9764-21bc72a07f60",
"metadata": {},
"source": [
"### Load the dataset"
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "1efffd3f-aece-4858-b615-8fb1f2997068",
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "b39d97f2fb434903b0521f5dea2fd37c",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Resolving data files: 0%| | 0/18 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "da3388e8f61e4aee98971ce520785c0c",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Resolving data files: 0%| | 0/18 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "db11101162ff455baed085d8c25a2c77",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Loading dataset shards: 0%| | 0/18 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from datasets import load_dataset\n",
"\n",
"base_dataset = load_dataset(\"ruslanmv/ai-medical-dataset\")\n",
"# define the train dataset as a random 80% of the data\n",
"train_dataset = base_dataset[\"train\"].train_test_split(test_size=0.2)[\"train\"]\n",
"# define the eval dataset as the remaining 20%\n",
"eval_dataset = base_dataset[\"train\"].train_test_split(test_size=0.2)[\"test\"]"
]
},
{
"cell_type": "markdown",
"id": "6be720c0-90a3-4ec0-bf1c-4d18e6b2bb31",
"metadata": {
"id": "Tu199rI3AMtG"
},
"source": [
"### Configure LoRA adapters"
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "86163110-f084-429e-bc70-5b281a679d1c",
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 67,
"referenced_widgets": [
"9f270a1faba44c1194d683a431d43f6f",
"7403ea0ef7f04e2389c351b0e10a68ce",
"fbc17ebe22854738b98bf2488c6105b1",
"22e04cb812df4f5ca621b041c213d17e",
"96d4f59ed68749eaaf9ad0c6a4e94f13",
"7a8bc62ce92a4571ad42ac2beef851ef",
"ae0117f952d94e469793bd8db3ad9a19",
"0b8f8a93515d49488642b008e47b4bf3",
"c9691e94467b4badbb71116136350c93",
"fcd773b4e6154af78e7d476c59294e25",
"7bd9e3fc12e64bc7bb95891dcd4153c7"
]
},
"executionInfo": {
"elapsed": 106706,
"status": "ok",
"timestamp": 1730998316179,
"user": {
"displayName": "Ronan Le Meillat",
"userId": "09161391957806824350"
},
"user_tz": -60
},
"id": "GTf6VxwwAMtG",
"outputId": "86a93163-6b30-4e8b-ca1c-7392922a4aa6"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"`low_cpu_mem_usage` was None, now default to True since model is quantized.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "591a95a8c90143a39a5616131c605ed3",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Loading checkpoint shards: 0%| | 0/5 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"import torch\n",
"from peft import LoraConfig, get_peft_model\n",
"from transformers import AutoProcessor, MllamaForConditionalGeneration, BitsAndBytesConfig, Idefics3ForConditionalGeneration\n",
"\n",
"DEVICE = \"cuda:0\"\n",
"USE_LORA = False\n",
"USE_QLORA = True\n",
"\n",
"processor = AutoProcessor.from_pretrained(\n",
" source_model_id,\n",
" do_image_splitting=False\n",
")\n",
"\n",
"if USE_QLORA or USE_LORA:\n",
" lora_config = LoraConfig(\n",
" r=8,\n",
" lora_alpha=8,\n",
" lora_dropout=0.1,\n",
" target_modules=[\"q_proj\", \"k_proj\", \"v_proj\", \"o_proj\",\n",
" \"gate_proj\", \"up_proj\", \"down_proj\",],\n",
" #target_modules='.*(text_model|modality_projection|perceiver_resampler).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$',\n",
" task_type=\"CAUSAL_LM\",\n",
" use_dora=False if USE_QLORA else True,\n",
" init_lora_weights=\"gaussian\"\n",
" )\n",
" if USE_QLORA:\n",
" bnb_config = BitsAndBytesConfig(\n",
" load_in_4bit=True,\n",
" bnb_4bit_quant_type=\"nf4\",\n",
" bnb_4bit_compute_dtype=torch.float16\n",
" )\n",
" model = MllamaForConditionalGeneration.from_pretrained(\n",
" source_model_id,\n",
" torch_dtype=torch.float16,\n",
" quantization_config=bnb_config if USE_QLORA else None,\n",
" )\n",
" model = get_peft_model(model, lora_config)\n",
" #model.add_adapter(lora_config)\n",
" #model.enable_adapters()\n",
"else:\n",
" model = MllamaForConditionalGeneration.from_pretrained(\n",
" source_model_id,\n",
" torch_dtype=torch.float16,\n",
" _attn_implementation=\"flash_attention_2\", # This works for A100 or H100\n",
" ).to(DEVICE)"
]
},
{
"cell_type": "code",
"execution_count": 17,
"id": "f2c2fd34-e1e4-427d-86d0-73bf74ff0005",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'question': 'What was the goal of the study?',\n",
" 'context': 'The aim of this study was to evaluate the success of steroid (PRED) withdrawal due to replacement by mycophenolate mofetil (MMF) in orthotopic liver transplant (OLT) recipients with autoimmune hepatitis (AIH). Thirty patients with AIH > 12 months after OLT randomized to receive either PRED and tacrolimus (TAC) or MMF and TAC were followed for 24 months. Withdrawal of steroids showed no difference regarding graft and patient survival. Also we demonstrated significantly lower glucose levels with lower HbA1c and a reduced need for insulin as well as a significantly lower serum cholesterol in the MMF group. Patients without steroids showed a lower incidence of osteopenia. Maintenance therapy in OLT patients with AIH may be performed safely using MMF instead of prednisone.'}"
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"eval_dataset\n",
"eval_dataset[24]"
]
},
{
"cell_type": "markdown",
"id": "01b0fe64-1f62-42a8-847d-9162f5015c4e",
"metadata": {
"id": "0JeaGZxHAMtG"
},
"source": [
"### Create Data Collator for IDEFICS3 format."
]
},
{
"cell_type": "code",
"execution_count": 57,
"id": "29d96aea-445d-482d-b7dc-861635a5389c",
"metadata": {
"executionInfo": {
"elapsed": 426,
"status": "ok",
"timestamp": 1730998596513,
"user": {
"displayName": "Ronan Le Meillat",
"userId": "09161391957806824350"
},
"user_tz": -60
},
"id": "X6TWyPHaAMtH"
},
"outputs": [],
"source": [
"class MyDataCollator:\n",
" def __init__(self, processor):\n",
" self.processor = processor\n",
" self.image_token_id = 128256\n",
"\n",
" def __call__(self, samples):\n",
" texts = []\n",
" images = []\n",
" for sample in samples:\n",
" question = sample[\"question\"]\n",
" answer = sample[\"context\"]\n",
" messages = [\n",
" {\n",
" \"role\": \"system\",\n",
" \"content\": [\n",
" {\"type\": \"text\", \"text\": prompt}\n",
" ]\n",
"\n",
" },\n",
" {\n",
" \"role\": \"user\",\n",
" \"content\": [\n",
" {\"type\": \"text\", \"text\": question },\n",
" ]\n",
" },\n",
" {\n",
" \"role\": \"assistant\",\n",
" \"content\": [\n",
" {\"type\": \"text\", \"text\": answer}\n",
" ]\n",
" }\n",
" ]\n",
" text = processor.apply_chat_template(messages, add_generation_prompt=False)\n",
" print(text)\n",
" texts.append(text.strip())\n",
"\n",
" batch = processor(text=texts, return_tensors=\"pt\", padding=True)\n",
"\n",
" labels = batch[\"input_ids\"].clone()\n",
" #labels[labels == processor.tokenizer.pad_token_id] = self.image_token_id\n",
" batch[\"labels\"] = labels\n",
"\n",
" return batch\n",
"\n",
"data_collator = MyDataCollator(processor)"
]
},
{
"cell_type": "markdown",
"id": "a8f2e613-5695-4558-9a13-66158d82bed9",
"metadata": {
"id": "vsq4TtIJAMtH"
},
"source": [
"### Setup training parameters"
]
},
{
"cell_type": "code",
"execution_count": 58,
"id": "f3cda658-05f6-4078-8d71-2d1c0352ecfa",
"metadata": {
"executionInfo": {
"elapsed": 1008,
"status": "ok",
"timestamp": 1730998601172,
"user": {
"displayName": "Ronan Le Meillat",
"userId": "09161391957806824350"
},
"user_tz": -60
},
"id": "Q_WKQFfoAMtH"
},
"outputs": [],
"source": [
"from transformers import TrainingArguments, Trainer\n",
"\n",
"training_args = TrainingArguments(\n",
" output_dir = output_dir,\n",
" overwrite_output_dir = False,\n",
" auto_find_batch_size = True,\n",
" learning_rate = 2e-4,\n",
" fp16 = True,\n",
" per_device_train_batch_size = 2,\n",
" per_device_eval_batch_size = 2,\n",
" gradient_accumulation_steps = 8,\n",
" dataloader_pin_memory = False,\n",
" save_total_limit = 3,\n",
" eval_strategy = \"steps\",\n",
" save_strategy = \"steps\",\n",
" eval_steps = 100,\n",
" save_steps = 10, # checkpoint each 10 steps\n",
" resume_from_checkpoint = True,\n",
" logging_steps = 5,\n",
" remove_unused_columns = False,\n",
" push_to_hub = True,\n",
" label_names = [\"labels\"],\n",
" load_best_model_at_end = False,\n",
" report_to = \"none\",\n",
" optim = \"paged_adamw_8bit\",\n",
" max_steps = 10, # remove this for training\n",
")"
]
},
{
"cell_type": "code",
"execution_count": 59,
"id": "e6569265-5941-4482-84e2-faf1b61b685c",
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"executionInfo": {
"elapsed": 426,
"status": "ok",
"timestamp": 1730998605441,
"user": {
"displayName": "Ronan Le Meillat",
"userId": "09161391957806824350"
},
"user_tz": -60
},
"id": "vSIo17mgAMtH",
"outputId": "3bebd35a-ed7f-49ee-e1bc-91594e8dcd24"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"max_steps is given, it will override any value given in num_train_epochs\n"
]
}
],
"source": [
"trainer = Trainer(\n",
" model = model,\n",
" args = training_args,\n",
" data_collator = data_collator,\n",
" train_dataset = train_dataset,\n",
" eval_dataset = eval_dataset,\n",
")"
]
},
{
"cell_type": "markdown",
"id": "916ac153-206b-488a-b783-3ad0c4ba21b6",
"metadata": {
"id": "pmlwDsOpAMtI"
},
"source": [
"### Start (or restart) Training"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "fb72a570-97e8-440e-b79f-640d8898e37c",
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "WQA84KnTAMtI",
"outputId": "ebb15160-f56e-4899-e608-b0d5fd0ba117"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the cross sensitivity with acetaminophen in aspirin-<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"Cross sensitivity with acetaminophen in aspirin-sensitive subjects with asthma.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the study of blood pressure control rate in Chinese hypertensive outpatients?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"[National survey of blood pressure control rate in Chinese hypertensive outpatients-China STATUS].<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the pH of the late secretory pathway?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"An autocatalytic cleavage in the C terminus of the human MUC2 mucin occurs at the low pH of the late secretory pathway.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the name of the uncontrolled diabetes mellitus mellitus mell<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"- Uncontrolled diabetes mellitus mellitus (as indicated by metabolic derangements and/or severe diabetes mellitus related uncontrolled organ complications)<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the name of the two unrelated yeast strains?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"The genome variability and meiotic gene expression patterns in two unrelated laboratory yeast strains, SK1 and W303, have been characterized using high-density oligonucleotide arrays. The statistical analysis and comparison of the data has allowed identification of: (1) genes with functional importance to meiosis and sporulation in yeast and (2) genes expressed in a strain-specific manner. The genome-wide data also reveal potential reasons why these strains display significant differences in the ability to make fertile spores. Molecularly tagged yeast deletion strains have been used to determine the contribution of each gene to the execution of the sporulation/germination pathway. The application of genetics and the new genomic technologies have allowed a quantum jump in our understanding of yeast molecular biology.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the name of the endproducts that are used in patients with systemic l<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"Accumulation of advanced glycation endproducts in patients with systemic lupus erythematosus.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the potential for developing small molecules antagonists?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"The development of specific targeted therapies, such as anti-TNF-alpha treatment, for chronic inflammatory disorders such as rheumatoid arthritis, has significantly improved treatment, although not all patients respond. Targeting cellular adhesion molecules and chemokines/chemokine receptors as regulators of the extravasation and migration of leukocytes may provide a novel approach for the treatment of these diseases. Moreover, the possibility of developing small-molecule antagonists offers an excellent method for the oral delivery of compounds with a short half-life.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the name of the type of cancer that is undergoing upper endoscopy?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"- People who are undergoing upper endoscopy as part of their medical care with a history of esophageal cancer, Barrett's esophagus, or upper gastrointestinal symptoms.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the purpose of the investigators' research?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"The investigators propose to find out how much Latino men know about the benefits and risks of prostate cancer screening and what factors influence how they decide whether or not to have screening. The investigators are also interested in learning how providing education information, both verbally and in print, about the risks and benefits of prostate cancer screening affects men's interest and knowledge.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the effectiveness of scalp cooling in preventing alopecia?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"The effectiveness of scalp cooling in preventing alopecia for patients receiving epirubicin and docetaxel.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"What is the PET/CT PET/CT PET/CT PET/CT PET/CT PET<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"68GA-BMV101 PET/CT IN PATIENTS WITH IDIOPATHIC PULMONARY FIBROSIS:<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"How many patients were enrolled in the MXF group?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"OBJECTIVE: To investigate whether adding moxifloxacin (MXF) to the standard anti-tuberculosis regimen can shorten the time to sputum culture conversion in pulmonary tuberculosis (PTB). METHODS: Adults with culture-positive PTB were divided into two treatment groups by their choice: standard regimen alone (HERZ group) and standard regimen plus daily 400 mg MXF in the first 2 months (MXF group). Sputum samples were collected thrice weekly in the first 8 weeks. The propensity score was calculated to estimate the conditional probability of entering the MXF group. Factors influencing time to culture conversion were investigated using Cox proportional hazards regression analysis stratified by propensity score. RESULTS: Sixty-two patients were enrolled in the MXF group and 88 in the HERZ group; respectively 51 and 72 completed the study. The regimen was modified before culture conversion in respectively 6 (12%) and 12 (16%; P = 0.47) patients, due to adverse effects. The time to culture conversion was shorter in the MXF group (HR 2.1, 95%CI 1.4-3.2). The culture conversion rate after 6 weeks of treatment was respectively 82% and 61% (P = 0.011, <0.05/4, calculated using the modified Bonferroni method). CONCLUSIONS: Adding MXF to the standard anti-tuberculosis regimen in the first 2 months was associated with a shorter time to culture conversion, a higher 6-week culture conversion rate and reduced transmission of tuberculosis.<|eot_id|>\n",
"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n",
"\n",
"Cutting Knowledge Date: December 2023\n",
"Today Date: 16 Nov 2024\n",
"\n",
"[{'type': 'text', 'text': 'You are a medical doctor with 15 year of experience verifying the knowledge of a new diploma medical doctor'}]<|eot_id|><|start_header_id|>user<|end_header_id|>\n",
"\n",
"Pyrimethamine.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
"\n",
"- Pyrimethamine.<|eot_id|>\n"
]
}
],
"source": [
"trainer.train()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "b109f2b9-f3cb-4732-8318-b74ed9e5aa25",
"metadata": {},
"outputs": [],
"source": []
}
],
"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": "3.12.7"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|