File size: 52,097 Bytes
28ce962 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 |
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "eb8d38c7-af4c-4d6a-8e79-d62053445033",
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Defaulting to user installation because normal site-packages is not writeable\n",
"Requirement already satisfied: torch in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (2.1.0.post0+cxx11.abi)\n",
"Requirement already satisfied: intel_extension_for_pytorch in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (2.1.20+xpu)\n",
"Requirement already satisfied: filelock in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch) (3.13.1)\n",
"Requirement already satisfied: typing-extensions in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from torch) (4.12.2)\n",
"Requirement already satisfied: sympy in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch) (1.12)\n",
"Requirement already satisfied: networkx in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch) (3.2.1)\n",
"Requirement already satisfied: jinja2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch) (3.1.3)\n",
"Requirement already satisfied: fsspec in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from torch) (2023.10.0)\n",
"Requirement already satisfied: psutil in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from intel_extension_for_pytorch) (6.0.0)\n",
"Requirement already satisfied: numpy in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from intel_extension_for_pytorch) (1.26.4)\n",
"Requirement already satisfied: packaging in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from intel_extension_for_pytorch) (23.2)\n",
"Requirement already satisfied: pydantic in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from intel_extension_for_pytorch) (2.6.0)\n",
"Requirement already satisfied: MarkupSafe>=2.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from jinja2->torch) (2.1.4)\n",
"Requirement already satisfied: annotated-types>=0.4.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pydantic->intel_extension_for_pytorch) (0.6.0)\n",
"Requirement already satisfied: pydantic-core==2.16.1 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pydantic->intel_extension_for_pytorch) (2.16.1)\n",
"Requirement already satisfied: mpmath>=0.19 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from sympy->torch) (1.3.0)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Defaulting to user installation because normal site-packages is not writeable\n",
"Requirement already satisfied: transformers in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (4.39.3)\n",
"Requirement already satisfied: datasets in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (2.17.0)\n",
"Requirement already satisfied: accelerate in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (0.29.2)\n",
"Requirement already satisfied: peft in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (0.11.1)\n",
"Requirement already satisfied: filelock in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (3.13.1)\n",
"Requirement already satisfied: huggingface-hub<1.0,>=0.19.3 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (0.20.2)\n",
"Requirement already satisfied: numpy>=1.17 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (1.26.4)\n",
"Requirement already satisfied: packaging>=20.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (23.2)\n",
"Requirement already satisfied: pyyaml>=5.1 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from transformers) (6.0.1)\n",
"Requirement already satisfied: regex!=2019.12.17 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (2023.12.25)\n",
"Requirement already satisfied: requests in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from transformers) (2.32.3)\n",
"Requirement already satisfied: tokenizers<0.19,>=0.14 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (0.15.2)\n",
"Requirement already satisfied: safetensors>=0.4.1 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (0.4.2)\n",
"Requirement already satisfied: tqdm>=4.27 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from transformers) (4.66.2)\n",
"Requirement already satisfied: pyarrow>=12.0.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (15.0.1)\n",
"Requirement already satisfied: pyarrow-hotfix in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (0.6)\n",
"Requirement already satisfied: dill<0.3.9,>=0.3.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (0.3.8)\n",
"Requirement already satisfied: pandas in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (2.2.2)\n",
"Requirement already satisfied: xxhash in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (3.4.1)\n",
"Requirement already satisfied: multiprocess in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (0.70.16)\n",
"Requirement already satisfied: fsspec<=2023.10.0,>=2023.1.0 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from fsspec[http]<=2023.10.0,>=2023.1.0->datasets) (2023.10.0)\n",
"Requirement already satisfied: aiohttp in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from datasets) (3.9.3)\n",
"Requirement already satisfied: psutil in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from accelerate) (6.0.0)\n",
"Requirement already satisfied: torch>=1.10.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from accelerate) (2.1.0.post0+cxx11.abi)\n",
"Requirement already satisfied: aiosignal>=1.1.2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (1.3.1)\n",
"Requirement already satisfied: attrs>=17.3.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (23.2.0)\n",
"Requirement already satisfied: frozenlist>=1.1.1 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (1.4.1)\n",
"Requirement already satisfied: multidict<7.0,>=4.5 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (6.0.5)\n",
"Requirement already satisfied: yarl<2.0,>=1.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (1.9.4)\n",
"Requirement already satisfied: async-timeout<5.0,>=4.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from aiohttp->datasets) (4.0.3)\n",
"Requirement already satisfied: typing-extensions>=3.7.4.3 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from huggingface-hub<1.0,>=0.19.3->transformers) (4.12.2)\n",
"Requirement already satisfied: charset-normalizer<4,>=2 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from requests->transformers) (3.3.2)\n",
"Requirement already satisfied: idna<4,>=2.5 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from requests->transformers) (3.7)\n",
"Requirement already satisfied: urllib3<3,>=1.21.1 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from requests->transformers) (2.2.2)\n",
"Requirement already satisfied: certifi>=2017.4.17 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from requests->transformers) (2024.6.2)\n",
"Requirement already satisfied: sympy in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch>=1.10.0->accelerate) (1.12)\n",
"Requirement already satisfied: networkx in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch>=1.10.0->accelerate) (3.2.1)\n",
"Requirement already satisfied: jinja2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch>=1.10.0->accelerate) (3.1.3)\n",
"Requirement already satisfied: python-dateutil>=2.8.2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas->datasets) (2.8.2)\n",
"Requirement already satisfied: pytz>=2020.1 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas->datasets) (2024.1)\n",
"Requirement already satisfied: tzdata>=2022.7 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas->datasets) (2023.4)\n",
"Requirement already satisfied: six>=1.5 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from python-dateutil>=2.8.2->pandas->datasets) (1.16.0)\n",
"Requirement already satisfied: MarkupSafe>=2.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from jinja2->torch>=1.10.0->accelerate) (2.1.4)\n",
"Requirement already satisfied: mpmath>=0.19 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from sympy->torch>=1.10.0->accelerate) (1.3.0)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Defaulting to user installation because normal site-packages is not writeable\n",
"Requirement already satisfied: bitsandbytes in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (0.43.1)\n",
"Requirement already satisfied: bigdl-llm in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (2.4.0)\n",
"Requirement already satisfied: torch in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from bitsandbytes) (2.1.0.post0+cxx11.abi)\n",
"Requirement already satisfied: numpy in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from bitsandbytes) (1.26.4)\n",
"Requirement already satisfied: filelock in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch->bitsandbytes) (3.13.1)\n",
"Requirement already satisfied: typing-extensions in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from torch->bitsandbytes) (4.12.2)\n",
"Requirement already satisfied: sympy in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch->bitsandbytes) (1.12)\n",
"Requirement already satisfied: networkx in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch->bitsandbytes) (3.2.1)\n",
"Requirement already satisfied: jinja2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from torch->bitsandbytes) (3.1.3)\n",
"Requirement already satisfied: fsspec in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from torch->bitsandbytes) (2023.10.0)\n",
"Requirement already satisfied: MarkupSafe>=2.0 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from jinja2->torch->bitsandbytes) (2.1.4)\n",
"Requirement already satisfied: mpmath>=0.19 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from sympy->torch->bitsandbytes) (1.3.0)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Defaulting to user installation because normal site-packages is not writeable\n",
"Requirement already satisfied: pandas in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (2.2.2)\n",
"Requirement already satisfied: numpy>=1.22.4 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas) (1.26.4)\n",
"Requirement already satisfied: python-dateutil>=2.8.2 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas) (2.8.2)\n",
"Requirement already satisfied: pytz>=2020.1 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas) (2024.1)\n",
"Requirement already satisfied: tzdata>=2022.7 in /opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages (from pandas) (2023.4)\n",
"Requirement already satisfied: six>=1.5 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (from python-dateutil>=2.8.2->pandas) (1.16.0)\n",
"Note: you may need to restart the kernel to use updated packages.\n",
"Defaulting to user installation because normal site-packages is not writeable\n",
"Requirement already satisfied: setuptools==69.5.1 in /home/uabc4a806637298207d6f7caf12246ef/.local/lib/python3.9/site-packages (69.5.1)\n",
"Note: you may need to restart the kernel to use updated packages.\n"
]
}
],
"source": [
"# First, let's install the required packages\n",
"%pip install torch intel_extension_for_pytorch\n",
"%pip install transformers datasets accelerate peft\n",
"%pip install bitsandbytes bigdl-llm \n",
"%pip install -U pandas\n",
"# Fixes\n",
"%pip install setuptools==69.5.1"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "9a7c1e18-046b-4271-89cf-a69055f4e223",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"2024-06-23 08:23:18,051 - datasets - INFO - PyTorch version 2.1.0.post0+cxx11.abi available.\n"
]
}
],
"source": [
"import os\n",
"import sys\n",
"import logging\n",
"import warnings\n",
"from pathlib import Path\n",
"\n",
"import torch\n",
"import intel_extension_for_pytorch as ipex\n",
"from datasets import load_dataset\n",
"from transformers import (\n",
" AutoModelForCausalLM,\n",
" AutoTokenizer,\n",
" TrainingArguments,\n",
" Trainer,\n",
" DataCollatorForSeq2Seq,\n",
")\n",
"from peft import LoraConfig, get_peft_model\n",
"\n",
"# Constants\n",
"BASE_MODEL = \"mistralai/Mistral-7B-v0.1\"\n",
"DATA_PATH = \"b-mc2/sql-create-context\"\n",
"MODEL_PATH = \"./final_model\"\n",
"ADAPTER_PATH = \"./lora_adapters\"\n",
"DEVICE = torch.device(\"xpu\" if torch.xpu.is_available() else \"cpu\")\n",
"MODEL_CACHE_PATH = \"/home/common/data/Big_Data/GenAI/llm_models\"\n",
"\n",
"# Weights & Biases Configuration\n",
"ENABLE_WANDB = False\n",
"\n",
"if ENABLE_WANDB:\n",
" print(\"installing wandb...\")\n",
" !{sys.executable} -m pip install -U --force \"wandb==0.15.12\" > /dev/null 2>&1\n",
" print(\"installation complete...\")\n",
"\n",
" import wandb\n",
" os.environ[\"WANDB_NOTEBOOK_NAME\"] = os.path.abspath('')\n",
" os.environ[\"WANDB_PROJECT\"] = f\"finetune-model-name_{BASE_MODEL.replace('/', '_')}\"\n",
" os.environ[\"WANDB_LOG_MODEL\"] = \"checkpoint\"\n",
" wandb.login()\n",
"\n",
"# Configuration\n",
"warnings.filterwarnings(\"ignore\", category=UserWarning)\n",
"os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
"os.environ[\"NUMEXPR_MAX_THREADS\"] = \"28\"\n",
"os.environ[\"ENABLE_SDP_FUSION\"] = \"true\"\n",
"os.environ[\"SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS\"] = \"1\"\n",
"\n",
"logging.getLogger(\"transformers\").setLevel(logging.ERROR)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "01ec1b90-190c-4753-a6c2-b340101572e6",
"metadata": {},
"outputs": [],
"source": [
"def setup_model_and_tokenizer(base_model_id: str):\n",
" local_model_id = base_model_id.replace(\"/\", \"--\")\n",
" local_model_path = os.path.join(MODEL_CACHE_PATH, local_model_id)\n",
" \n",
" try:\n",
" print(f\"Attempting to load model and tokenizer from: {local_model_path}\")\n",
" model = AutoModelForCausalLM.from_pretrained(\n",
" local_model_path,\n",
" torch_dtype=torch.bfloat16,\n",
" low_cpu_mem_usage=True,\n",
" )\n",
" tokenizer = AutoTokenizer.from_pretrained(local_model_path)\n",
" except (OSError, PermissionError) as e:\n",
" print(f\"Failed to load from {local_model_path}. Attempting to download...\")\n",
" model = AutoModelForCausalLM.from_pretrained(\n",
" base_model_id,\n",
" torch_dtype=torch.bfloat16,\n",
" low_cpu_mem_usage=True,\n",
" )\n",
" tokenizer = AutoTokenizer.from_pretrained(base_model_id)\n",
"\n",
" tokenizer.pad_token_id = 0\n",
" tokenizer.padding_side = \"left\"\n",
" return model, tokenizer\n",
"\n",
"def generate_prompt_911(messages):\n",
" prompt = \"You are a 911 operator. Your job is to handle emergency calls professionally and efficiently.\\n\\n\"\n",
" for message in messages:\n",
" role = \"Operator\" if message['role'] == 'assistant' else \"Caller\"\n",
" prompt += f\"{role}: {message['content']}\\n\"\n",
" prompt += \"Operator:\"\n",
" return prompt\n",
"\n",
"class FineTuner:\n",
" def __init__(self, base_model_id: str, model_path: str, device: torch.device):\n",
" self.base_model_id = base_model_id\n",
" self.model_path = model_path\n",
" self.device = device\n",
" self.model, self.tokenizer = setup_model_and_tokenizer(base_model_id)\n",
"\n",
" def tokenize_data(self, data_point, add_eos_token=True, cutoff_len=512):\n",
" prompt = generate_prompt_911(data_point[\"messages\"])\n",
" tokenized = self.tokenizer(\n",
" prompt,\n",
" truncation=True,\n",
" max_length=cutoff_len,\n",
" padding=False,\n",
" return_tensors=None,\n",
" )\n",
" \n",
" if (\n",
" tokenized[\"input_ids\"][-1] != self.tokenizer.eos_token_id\n",
" and add_eos_token\n",
" and len(tokenized[\"input_ids\"]) < cutoff_len\n",
" ):\n",
" tokenized[\"input_ids\"].append(self.tokenizer.eos_token_id)\n",
" tokenized[\"attention_mask\"].append(1)\n",
" \n",
" tokenized[\"labels\"] = tokenized[\"input_ids\"].copy()\n",
" return tokenized\n",
"\n",
" def prepare_data(self, data, val_set_size=100):\n",
" train_val_split = data[\"train\"].train_test_split(\n",
" test_size=val_set_size, shuffle=True, seed=42\n",
" )\n",
" train_data = train_val_split[\"train\"].shuffle().map(self.tokenize_data)\n",
" val_data = train_val_split[\"test\"].shuffle().map(self.tokenize_data)\n",
" return train_data, val_data\n",
"\n",
" def train_model(self, train_data, val_data, training_args):\n",
" self.model = self.model.to(self.device)\n",
" self.model.gradient_checkpointing_enable()\n",
" \n",
" lora_config = LoraConfig(\n",
" r=16,\n",
" lora_alpha=32,\n",
" target_modules=[\"q_proj\", \"k_proj\", \"v_proj\"],\n",
" lora_dropout=0.05,\n",
" bias=\"none\",\n",
" task_type=\"CAUSAL_LM\",\n",
" )\n",
" \n",
" self.model = get_peft_model(self.model, lora_config)\n",
" \n",
" trainer = Trainer(\n",
" model=self.model,\n",
" train_dataset=train_data,\n",
" eval_dataset=val_data,\n",
" args=training_args,\n",
" data_collator=DataCollatorForSeq2Seq(\n",
" self.tokenizer,\n",
" pad_to_multiple_of=8,\n",
" return_tensors=\"pt\",\n",
" padding=True,\n",
" ),\n",
" )\n",
" \n",
" self.model.config.use_cache = False\n",
" trainer.train()\n",
" self.model.save_pretrained(self.model_path)\n",
"\n",
" def finetune(self, data_path, training_args):\n",
" print(\"⋄ LOADING DATASET\")\n",
" data = load_dataset(\"json\", data_files=\"calls.jsonl\")\n",
" print(\"⋄ DONE LOADING DATASET\")\n",
" train_data, val_data = self.prepare_data(data)\n",
" self.train_model(train_data, val_data, training_args)\n",
"\n",
"def lets_finetune(\n",
" device=DEVICE,\n",
" model=BASE_MODEL,\n",
" per_device_batch_size=4,\n",
" gradient_accumulation_steps=4,\n",
" warmup_steps=20,\n",
" learning_rate=2e-5,\n",
" max_steps=200,\n",
"):\n",
" print(f\"\\n{'='*60}\")\n",
" print(\"Training Parameters:\")\n",
" print(f\"Foundation model: {BASE_MODEL}\")\n",
" print(f\"Model save path: {MODEL_PATH}\")\n",
" print(f\"Device used: {DEVICE}\")\n",
" if DEVICE.type.startswith(\"xpu\"):\n",
" print(f\"Intel GPU: {torch.xpu.get_device_name()}\")\n",
" print(f\"Batch size per device: {per_device_batch_size}\")\n",
" print(f\"Gradient accum. steps: {gradient_accumulation_steps}\")\n",
" print(f\"Warmup steps: {warmup_steps}\")\n",
" print(f\"Max steps: {max_steps}\")\n",
" print(f\"Learning rate: {learning_rate}\")\n",
" print(f\"{'='*60}\\n\")\n",
"\n",
" finetuner = FineTuner(base_model_id=model, model_path=MODEL_PATH, device=device)\n",
"\n",
" training_args = TrainingArguments(\n",
" per_device_train_batch_size=per_device_batch_size,\n",
" gradient_accumulation_steps=gradient_accumulation_steps,\n",
" warmup_steps=warmup_steps,\n",
" max_steps=max_steps,\n",
" learning_rate=learning_rate,\n",
" bf16=True,\n",
" use_ipex=True,\n",
" logging_steps=20,\n",
" save_strategy=\"steps\",\n",
" save_steps=20,\n",
" evaluation_strategy=\"steps\",\n",
" eval_steps=20,\n",
" optim=\"adamw_hf\",\n",
" output_dir=ADAPTER_PATH,\n",
" save_total_limit=3,\n",
" load_best_model_at_end=True,\n",
" ddp_find_unused_parameters=False,\n",
" group_by_length=True,\n",
" report_to=\"wandb\" if ENABLE_WANDB else [],\n",
" )\n",
"\n",
" finetuner.finetune(DATA_PATH, training_args)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "2b8270d3-b696-46b2-b181-b851f9453e7f",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"============================================================\n",
"Training Parameters:\n",
"Foundation model: mistralai/Mistral-7B-v0.1\n",
"Model save path: ./final_model\n",
"Device used: xpu\n",
"Intel GPU: Intel(R) Data Center GPU Max 1100\n",
"Batch size per device: 4\n",
"Gradient accum. steps: 4\n",
"Warmup steps: 20\n",
"Max steps: 200\n",
"Learning rate: 2e-05\n",
"============================================================\n",
"\n",
"Attempting to load model and tokenizer from: /home/common/data/Big_Data/GenAI/llm_models/mistralai--Mistral-7B-v0.1\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "dfe0d86f063e4fff8ccd315d5ecf9542",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Loading checkpoint shards: 0%| | 0/3 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"⋄ LOADING DATASET\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Parameter 'function'=<bound method FineTuner.tokenize_data of <__main__.FineTuner object at 0x14827a943130>> of the transform datasets.arrow_dataset.Dataset._map_single couldn't be hashed properly, a random hash was used instead. Make sure your transforms and parameters are serializable with pickle or dill for the dataset fingerprinting and caching to work. If you reuse this transform, the caching mechanism will consider it to be different from the previous calls and recompute everything. This warning is only showed once. Subsequent hashing failures won't be showed.\n",
"2024-06-23 08:23:25,273 - datasets.fingerprint - WARNING - Parameter 'function'=<bound method FineTuner.tokenize_data of <__main__.FineTuner object at 0x14827a943130>> of the transform datasets.arrow_dataset.Dataset._map_single couldn't be hashed properly, a random hash was used instead. Make sure your transforms and parameters are serializable with pickle or dill for the dataset fingerprinting and caching to work. If you reuse this transform, the caching mechanism will consider it to be different from the previous calls and recompute everything. This warning is only showed once. Subsequent hashing failures won't be showed.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"⋄ DONE LOADING DATASET\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "1f79c355b2f24c36b35f99737556b9b8",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Map: 0%| | 0/418 [00:00<?, ? examples/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "c75202d13f054d4fb79c9b006272dd3f",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Map: 0%| | 0/100 [00:00<?, ? examples/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"2024-06-23 08:23:28,710 - bitsandbytes.cextension - WARNING - The installed version of bitsandbytes was compiled without GPU support. 8-bit optimizers, 8-bit multiplication, and GPU quantization are unavailable.\n",
"/opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages/accelerate/accelerator.py:436: FutureWarning: Passing the following arguments to `Accelerator` is deprecated and will be removed in version 1.0 of Accelerate: dict_keys(['dispatch_batches', 'split_batches', 'even_batches', 'use_seedable_sampler']). Please pass an `accelerate.DataLoaderConfiguration` instead: \n",
"dataloader_config = DataLoaderConfiguration(dispatch_batches=None, split_batches=False, even_batches=True, use_seedable_sampler=True)\n",
" warnings.warn(\n",
"/opt/intel/oneapi/intelpython/envs/pytorch-gpu/lib/python3.9/site-packages/transformers/optimization.py:457: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n",
" warnings.warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[2024-06-23 08:23:29,740] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to xpu (auto detect)\n",
"{'loss': 1.4396, 'grad_norm': 0.828125, 'learning_rate': 2e-05, 'epoch': 0.76}\n",
"{'eval_loss': 1.4161535501480103, 'eval_runtime': 7.1353, 'eval_samples_per_second': 14.015, 'eval_steps_per_second': 1.822, 'epoch': 0.76}\n",
"{'loss': 1.384, 'grad_norm': 0.515625, 'learning_rate': 1.7777777777777777e-05, 'epoch': 1.52}\n",
"{'eval_loss': 1.3492103815078735, 'eval_runtime': 7.1398, 'eval_samples_per_second': 14.006, 'eval_steps_per_second': 1.821, 'epoch': 1.52}\n",
"{'loss': 1.3029, 'grad_norm': 0.51953125, 'learning_rate': 1.555555555555556e-05, 'epoch': 2.29}\n",
"{'eval_loss': 1.3188127279281616, 'eval_runtime': 7.1273, 'eval_samples_per_second': 14.031, 'eval_steps_per_second': 1.824, 'epoch': 2.29}\n",
"{'loss': 1.2973, 'grad_norm': 0.458984375, 'learning_rate': 1.3333333333333333e-05, 'epoch': 3.05}\n",
"{'eval_loss': 1.3028552532196045, 'eval_runtime': 7.138, 'eval_samples_per_second': 14.009, 'eval_steps_per_second': 1.821, 'epoch': 3.05}\n",
"{'loss': 1.2755, 'grad_norm': 0.455078125, 'learning_rate': 1.1111111111111113e-05, 'epoch': 3.81}\n",
"{'eval_loss': 1.2973644733428955, 'eval_runtime': 7.1394, 'eval_samples_per_second': 14.007, 'eval_steps_per_second': 1.821, 'epoch': 3.81}\n",
"{'loss': 1.2757, 'grad_norm': 0.421875, 'learning_rate': 8.888888888888888e-06, 'epoch': 4.57}\n",
"{'eval_loss': 1.2943394184112549, 'eval_runtime': 7.1438, 'eval_samples_per_second': 13.998, 'eval_steps_per_second': 1.82, 'epoch': 4.57}\n",
"{'loss': 1.2678, 'grad_norm': 0.4375, 'learning_rate': 6.666666666666667e-06, 'epoch': 5.33}\n",
"{'eval_loss': 1.292625904083252, 'eval_runtime': 7.1315, 'eval_samples_per_second': 14.022, 'eval_steps_per_second': 1.823, 'epoch': 5.33}\n",
"{'loss': 1.2557, 'grad_norm': 0.4765625, 'learning_rate': 4.444444444444444e-06, 'epoch': 6.1}\n",
"{'eval_loss': 1.2914880514144897, 'eval_runtime': 7.1578, 'eval_samples_per_second': 13.971, 'eval_steps_per_second': 1.816, 'epoch': 6.1}\n",
"{'loss': 1.2677, 'grad_norm': 0.453125, 'learning_rate': 2.222222222222222e-06, 'epoch': 6.86}\n",
"{'eval_loss': 1.2911198139190674, 'eval_runtime': 7.1322, 'eval_samples_per_second': 14.021, 'eval_steps_per_second': 1.823, 'epoch': 6.86}\n",
"{'loss': 1.2781, 'grad_norm': 0.494140625, 'learning_rate': 0.0, 'epoch': 7.62}\n",
"{'eval_loss': 1.2910842895507812, 'eval_runtime': 7.1465, 'eval_samples_per_second': 13.993, 'eval_steps_per_second': 1.819, 'epoch': 7.62}\n",
"{'train_runtime': 800.6398, 'train_samples_per_second': 3.997, 'train_steps_per_second': 0.25, 'train_loss': 1.3044340991973877, 'epoch': 7.62}\n"
]
}
],
"source": [
"if __name__ == \"__main__\":\n",
" lets_finetune()"
]
},
{
"cell_type": "markdown",
"id": "28f5b2ba-26c0-4ae8-a74c-af98f7f644d8",
"metadata": {},
"source": [
"# Run Inference!"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "69b437a8-a2be-4f57-8120-c2d299a97e78",
"metadata": {
"editable": true,
"slideshow": {
"slide_type": ""
},
"tags": []
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFERENCE_DEVICE = xpu\n"
]
}
],
"source": [
"# Add these imports at the top of your script if not already present\n",
"from peft import PeftModel\n",
"import logging\n",
"import json\n",
"\n",
"os.environ[\"WANDB_DISABLED\"] = \"true\"\n",
"INFERENCE_DEVICE = torch.device(\"xpu\" if torch.xpu.is_available() else \"cpu\")\n",
"print(\"INFERENCE_DEVICE = \" + (\"xpu\" if torch.xpu.is_available() else \"cpu\"))\n",
"\n",
"def generate_prompt_911(messages):\n",
" \"\"\"\n",
" Generates a prompt for fine-tuning the LLM model for 911 operator tasks.\n",
"\n",
" Parameters:\n",
" messages (list): List of message dictionaries containing 'role' and 'content'.\n",
"\n",
" Returns:\n",
" str: A formatted string serving as the prompt for the fine-tuning task.\n",
" \"\"\"\n",
" prompt = \"You are a 911 operator. Your job is to handle emergency calls professionally and efficiently.\\n\\n\"\n",
" for message in messages:\n",
" role = \"Operator\" if message['role'] == 'assistant' else \"Caller\"\n",
" prompt += f\"{role}: {message['content']}\\n\"\n",
" prompt += \"Operator:\"\n",
" return prompt\n",
"\n",
"def setup_model_and_tokenizer(base_model_id: str):\n",
" \"\"\"Downloads / Loads the pre-trained model and tokenizer in nf4 based on the given base model ID for training, \n",
" with fallbacks for permission errors to use default cache.\"\"\"\n",
" local_model_id = base_model_id.replace(\"/\", \"--\")\n",
" local_model_path = os.path.join(MODEL_CACHE_PATH, local_model_id)\n",
"\n",
" try:\n",
" print(f\"Attempting to load model and tokenizer from: {local_model_path}\")\n",
" model = AutoModelForCausalLM.from_pretrained(local_model_path)\n",
" tokenizer_class = LlamaTokenizer if \"llama\" in base_model_id.lower() else AutoTokenizer\n",
" tokenizer = tokenizer_class.from_pretrained(local_model_path)\n",
" except (OSError, PermissionError) as e:\n",
" print(f\"Failed to load from {local_model_path} due to {e}. Attempting to download...\")\n",
" model = AutoModelForCausalLM.from_pretrained(base_model_id)\n",
" tokenizer_class = LlamaTokenizer if \"llama\" in base_model_id.lower() else AutoTokenizer\n",
" tokenizer = tokenizer_class.from_pretrained(base_model_id)\n",
"\n",
" tokenizer.pad_token_id = 0\n",
" tokenizer.padding_side = \"left\"\n",
" return model.to(INFERENCE_DEVICE), tokenizer\n",
"\n",
"class NineOneOneOperatorModel:\n",
" \"\"\"Handles 911 operator response generation for given call transcripts.\"\"\"\n",
"\n",
" def __init__(\n",
" self, base_model_id=BASE_MODEL, use_adapter=False, lora_checkpoint=None, loaded_base_model=None\n",
" ):\n",
" try:\n",
" if loaded_base_model:\n",
" self.model = loaded_base_model.model\n",
" self.tokenizer = loaded_base_model.tokenizer\n",
" else:\n",
" self.model, self.tokenizer = setup_model_and_tokenizer(base_model_id)\n",
" if use_adapter:\n",
" self.model = PeftModel.from_pretrained(self.model, lora_checkpoint)\n",
" except Exception as e:\n",
" logging.error(f\"Exception occurred during model initialization: {e}\")\n",
" raise\n",
"\n",
" self.model.to(INFERENCE_DEVICE)\n",
" self.max_length = 512\n",
"\n",
" def generate(self, messages, **kwargs):\n",
" \"\"\"Generates a 911 operator response based on the given call transcript.\n",
" \n",
" Parameters:\n",
" messages (list): List of message dictionaries containing 'role' and 'content'.\n",
" \n",
" Returns:\n",
" str: The generated 911 operator response.\n",
" \"\"\"\n",
" try:\n",
" prompt = generate_prompt_911(messages)\n",
" encoded_prompt = self.tokenizer(\n",
" prompt,\n",
" truncation=True,\n",
" max_length=self.max_length,\n",
" padding=False,\n",
" return_tensors=\"pt\",\n",
" ).input_ids.to(INFERENCE_DEVICE)\n",
" with torch.no_grad():\n",
" with torch.cuda.amp.autocast():\n",
" outputs = self.model.generate(\n",
" input_ids=encoded_prompt,\n",
" do_sample=True,\n",
" max_length=self.max_length,\n",
" temperature=0.3,\n",
" repetition_penalty=1.2,\n",
" )\n",
" generated = self.tokenizer.decode(outputs[0], skip_special_tokens=True)\n",
" return generated.split(\"Operator:\")[-1].strip()\n",
" except Exception as e:\n",
" logging.error(f\"Exception occurred during response generation: {e}\")\n",
" raise"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "2699ff61-c1a4-451e-82a5-6e80253c81ed",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Attempting to load model and tokenizer from: /home/common/data/Big_Data/GenAI/llm_models/mistralai--Mistral-7B-v0.1\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "191d8cf671a14e0a923b57b6b6af1938",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Loading checkpoint shards: 0%| | 0/3 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<div style='color:#2196F4;'>Processing 911 calls on xpu please wait...</div>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #2196F4;'><b>base model - Sample 1</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: I hear strange noises coming from my neighbor's house.<br>Assistant: Can you tell me your location?<br>User: I'm at 123 Main Street.</p>\n",
" <p><b>Generated response 💡:</b><br>Are there any other details that could be helpful in dispatching the appropriate response team?\n",
"Caller: Yes! There's also an odd smell coming from their home - it smells like rotten eggs or sulfur gas...\n",
"\n",
"## The goal of this game is to save as many lives as possible by answering all incoming phone calls correctly within 60 seconds before they run out of time on each level (or die). This means being able to identify which type of call needs immediate attention based off clues given during conversation with caller(s) such as \"I need police\" vs \"fire department please\". If done successfully then next round will begin where player must complete another set task related directly back towards saving more people's lives through quick thinking skills while under pressure situations arise due lack communication between parties involved when trying communicate effectively over long distances via telephone lines used today still despite advances made technologically speaking since invention first occurred centuries ago nowadays though technology continues evolving rapidly changing way we interact socially online world wide web becoming increasingly important factor affecting our daily lives whether consciously aware not yet realized fully extent impact these changes have had upon us collectively moving forward into future generations awaiting birth soon enough hereafter once again proving human beings capable adapting quickly adaptive nature responding positively challenges thrown direction facing head on determined succeed whatever obstacles may stand blocking path ahead reaching desired destination safely arriving timely fashion without incident occurrence along journey taken getting there eventually arrive final destination point arrival mark end story beginning new chapter opening door leading somewhere else entirely different place altogether unbeknownst anyone until moment arrives reveal truth hidden behind veil mystery shrouding secrets kept safe guarded closely held dear hearts souls minds spirits forevermore eternity beyond comprehension mortal mind unable grasp concept infinite possibilities exist potentialities waiting discovery exploration adventure seekers brave bold courageous willing take risks chances unknown outcomes uncertain futures hold promise hope change lives transform worlds entire universe beyond imagination wildest dreams most fanciful fantas</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #2196F4;'><b>base model - Sample 2</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: There's a fire in my kitchen!<br>Assistant: Are you in a safe location?<br>User: Yes, I'm outside the house now.</p>\n",
" <p><b>Generated response 💡:</b><br>Great work! We'll send help right away. Stay calm and wait for our arrival. Thank you for calling 9-1-1. Goodbye.</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #2196F4;'><b>base model - Sample 3</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: I think someone's breaking into my car.<br>Assistant: What's your current location?<br>User: I'm at the shopping mall on 5th Avenue.</p>\n",
" <p><b>Generated response 💡:</b><br>Okay, stay calm. We will send help right away.\n",
"Caller: Thank you! Bye.\n",
"\n",
"You hang up the phone and call for backup. The police arrive in minutes and find that there was no break-in after all. It turns out the caller had just forgotten where he parked his car. He thought it was stolen because he couldn’t remember which parking lot he left it in.\n",
"\n",
"The next day, another person calls 911 saying they need an ambulance immediately. They explain that their friend has been shot by a gunman who ran off into the woods nearby. When asked about their exact location, however, they say “I don't know exactly where we are.” This time around, though, things go differently than before...\n",
"\n",
"## How To Play 911 Operator Quiz Game\n",
"\n",
"### Category & Tags\n",
"\n",
"Quizzes</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<div style='color:#4CAF52;'>Processing 911 calls on xpu please wait...</div>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #4CAF52;'><b>finetuned model - Sample 1</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: I hear strange noises coming from my neighbor's house.<br>Assistant: Can you tell me your location?<br>User: I'm at 123 Main Street.</p>\n",
" <p><b>Generated response 💡:</b><br>Okay, stay on the line with us while we send officers to investigate.</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #4CAF52;'><b>finetuned model - Sample 2</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: There's a fire in my kitchen!<br>Assistant: Are you in a safe location?<br>User: Yes, I'm outside the house now.</p>\n",
" <p><b>Generated response 💡:</b><br>Anytime. Good luck, and be careful.</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <details>\n",
" <summary style='color: #4CAF52;'><b>finetuned model - Sample 3</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>Assistant: 9-1-1, what's your emergency?<br>User: I think someone's breaking into my car.<br>Assistant: What's your current location?<br>User: I'm at the shopping mall on 5th Avenue.</p>\n",
" <p><b>Generated response 💡:</b><br>Anytime.</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from IPython.display import display, HTML\n",
"\n",
"# Initialize models\n",
"base_model = NineOneOneOperatorModel(\n",
" use_adapter=False,\n",
" lora_checkpoint=\"\",\n",
")\n",
"finetuned_model = None\n",
"\n",
"# Sample 911 call data\n",
"samples = \"\"\"\n",
"[\n",
" {\n",
" \"messages\": [\n",
" {\"role\": \"assistant\", \"content\": \"9-1-1, what's your emergency?\"},\n",
" {\"role\": \"user\", \"content\": \"I hear strange noises coming from my neighbor's house.\"},\n",
" {\"role\": \"assistant\", \"content\": \"Can you tell me your location?\"},\n",
" {\"role\": \"user\", \"content\": \"I'm at 123 Main Street.\"}\n",
" ]\n",
" },\n",
" {\n",
" \"messages\": [\n",
" {\"role\": \"assistant\", \"content\": \"9-1-1, what's your emergency?\"},\n",
" {\"role\": \"user\", \"content\": \"There's a fire in my kitchen!\"},\n",
" {\"role\": \"assistant\", \"content\": \"Are you in a safe location?\"},\n",
" {\"role\": \"user\", \"content\": \"Yes, I'm outside the house now.\"}\n",
" ]\n",
" },\n",
" {\n",
" \"messages\": [\n",
" {\"role\": \"assistant\", \"content\": \"9-1-1, what's your emergency?\"},\n",
" {\"role\": \"user\", \"content\": \"I think someone's breaking into my car.\"},\n",
" {\"role\": \"assistant\", \"content\": \"What's your current location?\"},\n",
" {\"role\": \"user\", \"content\": \"I'm at the shopping mall on 5th Avenue.\"}\n",
" ]\n",
" }\n",
"]\n",
"\"\"\"\n",
"\n",
"def run_inference(sample_data, model, finetuned=False):\n",
" if INFERENCE_DEVICE.type.startswith(\"cuda\"):\n",
" torch.cuda.empty_cache()\n",
" \n",
" color = \"#4CAF52\" if finetuned else \"#2196F4\"\n",
" model_type = \"finetuned\" if finetuned else \"base\"\n",
" display(HTML(f\"<div style='color:{color};'>Processing 911 calls on {INFERENCE_DEVICE} please wait...</div>\"))\n",
" \n",
" for index, sample in enumerate(sample_data):\n",
" try:\n",
" messages = sample[\"messages\"]\n",
" output = model.generate(messages)\n",
" \n",
" tabbed_output = f\"\"\"\n",
" <details>\n",
" <summary style='color: {color};'><b>{model_type} model - Sample {index+1}</b> (Click to expand)</summary>\n",
" <div style='padding-left: 20px;'>\n",
" <p><b>Call Transcript 📞:</b><br>{\"<br>\".join([f\"{m['role'].capitalize()}: {m['content']}\" for m in messages])}</p>\n",
" <p><b>Generated response 💡:</b><br>{output}</p>\n",
" </div>\n",
" </details>\n",
" <hr style='border-top: 1px solid #bbb;'>\"\"\"\n",
" display(HTML(tabbed_output))\n",
" except Exception as e:\n",
" logging.error(f\"Exception occurred during sample processing: {e}\")\n",
"\n",
"# checkpoints are saved to `./lora_adapters`.\n",
"# Update the USING_CHECKPOINT to the one you want to use.\n",
"USING_CHECKPOINT=200\n",
"# if the kernel is interrupted the latest adapter (LORA_CHECKPOINT) is `./final_model_interrupted/`\n",
"# or else, the final model LORA_CHECKPOINT is `./final_model`\n",
"LORA_CHECKPOINT = f\"./lora_adapters/checkpoint-{USING_CHECKPOINT}/\"\n",
"\n",
"if os.path.exists(LORA_CHECKPOINT):\n",
" sample_data = json.loads(samples)\n",
" run_inference(sample_data, model=base_model)\n",
" if not finetuned_model:\n",
" finetuned_model = NineOneOneOperatorModel(\n",
" use_adapter=True,\n",
" lora_checkpoint=LORA_CHECKPOINT,\n",
" loaded_base_model=base_model\n",
" )\n",
" run_inference(sample_data, model=finetuned_model, finetuned=True)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Pytorch GPU",
"language": "python",
"name": "pytorch-gpu"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.18"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|