Spaces:
Runtime error
Runtime error
File size: 34,212 Bytes
9de5882 |
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 |
{
"cells": [
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"This Notebook is to test the various OpenAI models, prompts, and number of few-shot examples to see how they perform on the same task."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"!pip install wandb --upgrade openai datasets -qU"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"from dotenv import load_dotenv\n",
"load_dotenv()\n",
"\n",
"import openai\n",
"\n",
"# set OPENAI_API_KEY environment variable from .env file\n",
"openai.api_key = os.getenv(\"OPENAI_API_KEY\")\n",
"\n",
"# import OpenAIChatCompletions class from openai_chat_completion.py file and compare_completion_and_prediction function from util.py file\n",
"from openai_chat_completion import OpenAIChatCompletions\n",
"from util import compare_completion_and_prediction"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"Models:\n",
"- gpt-3.5-turbo\n",
"- gpt-4\n",
"\n",
"Prompts:\n",
"- gpt4-system-message.txt\n",
"\n",
"Few-shot examples:\n",
"> 0 ... 10"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"wandb setup:\n",
"- entity: kaleidoscope-data\n",
"- project: cookies_llm_experimental_eval\n",
"- tags: gpt-3.5-turbo, gpt-4, gpt4-system-message, few-shot"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Failed to detect the name of this notebook, you can set it manually with the WANDB_NOTEBOOK_NAME environment variable to enable code saving.\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: Logging into wandb.ai. (Learn how to deploy a W&B server locally: https://wandb.me/wandb-server)\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: You can find your API key in your browser here: https://wandb.ai/authorize\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: Paste an API key from your profile and hit enter, or press ctrl+c to quit:\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[32m\u001b[41mERROR\u001b[0m API key must be 40 characters long, yours was 48\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: Logging into wandb.ai. (Learn how to deploy a W&B server locally: https://wandb.me/wandb-server)\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: You can find your API key in your browser here: https://wandb.ai/authorize\n",
"\u001b[34m\u001b[1mwandb\u001b[0m: Paste an API key from your profile and hit enter, or press ctrl+c to quit:\u001b[34m\u001b[1mwandb\u001b[0m: Appending key for api.wandb.ai to your netrc file: /home/cmagganas/.netrc\n"
]
},
{
"data": {
"text/html": [
"Tracking run with wandb version 0.15.4"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Run data is saved locally in <code>/home/cmagganas/kaleidoscope/llm_data_cleaner/app/wandb/run-20230626_114056-rbtf91s6</code>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Syncing run <strong><a href='https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval/runs/rbtf91s6' target=\"_blank\">rose-puddle-7</a></strong> to <a href='https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
" View project at <a href='https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval' target=\"_blank\">https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval</a>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
" View run at <a href='https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval/runs/rbtf91s6' target=\"_blank\">https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval/runs/rbtf91s6</a>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from wandb.integration.openai import autolog\n",
"\n",
"autolog({\"project\":\"cookies_llm_experimental_eval\",\n",
" \"entity\": \"kaleidoscope-data\",\n",
" \"group\": \"cookies\",\n",
" \"job_type\": \"eval\"})"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"# create an empty dataframe to store predictions\n",
"import pandas as pd\n",
"predictions_df = pd.DataFrame(columns=['model', 'system_message', 'n_shot', 'prompt', 'completion', 'prediction'])\n",
"\n",
"models_to_test = [\"gpt-4\", \"gpt-3.5-turbo\"]\n",
"sys_mes_to_test = [\"../prompts/gpt4-system-message.txt\", \"../prompts/gpt4-system-message2.txt\"] # names are arbitrary, same prompts but with \"####\" in system message 2\n",
"n_shots_to_test = [None, 1, 2, 3, 5]"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"# if rerunning the below cell is required, set the following to True\n",
"rerun = False\n",
"if rerun:\n",
" predictions_df = pd.read_csv('../data/cookies_llm_eval_predictions.csv')"
]
},
{
"cell_type": "code",
"execution_count": 178,
"metadata": {},
"outputs": [],
"source": [
"# get predictions for all combinations of models, prompts, and n_shot values\n",
"# save predictions to dataframe and then to csv in data folder after each iteration\n",
"\n",
"# loop through models_to_test\n",
"for model in models_to_test:\n",
" # loop through prompts_to_test\n",
" for system_message in sys_mes_to_test:\n",
" # instantiate OpenAIChatCompletions class\n",
" chat = OpenAIChatCompletions(model=model, system_message=system_message)\n",
" # loop through n_shots_to_test\n",
" for n_shot in n_shots_to_test:\n",
" sys_mes_var = 1 if system_message == \"../prompts/gpt4-system-message.txt\" else 2\n",
" n_shot_var = 0 if n_shot == None else n_shot\n",
" # check if predictions for this model, system_message, and n_shot value have already been made\n",
" if predictions_df[(predictions_df['model'] == model) & (predictions_df['system_message'] == sys_mes_var) & (predictions_df['n_shot'] == n_shot_var)].shape[0] == 0:\n",
" prompts, completions, predictions = chat.predict_jsonl(n_shot=n_shot)\n",
" else:\n",
" # skip if predictions for this model, system_message, and n_shot value have already been made\n",
" continue\n",
" # save predictions to dataframe\n",
" df_to_append = pd.DataFrame({'model': model, 'system_message': sys_mes_var, 'n_shot': n_shot_var, 'prompt': prompts, 'completion': completions, 'prediction': predictions})\n",
" df_right = df_to_append['prediction'].apply(pd.Series)\n",
" df_right['prediction'] = df_right['choices'].apply(lambda x: x[0]['message']['content']).drop(columns=['choices'])\n",
" df_to_append = pd.concat([df_to_append[['model', 'system_message', 'n_shot', 'prompt', 'completion']], df_right], axis=1)\n",
" df_to_append.columns = ['model', 'system_message', 'n_shot', 'prompt', 'completion', 'id', 'object', 'created', 'openai_model', 'choices', 'usage', 'prediction']\n",
" # save predictions to dataframe\n",
" predictions_df = pd.concat([predictions_df, df_to_append], ignore_index=True)\n",
" # delete duplicates from dataframe\n",
" predictions_df = predictions_df[~predictions_df.duplicated(subset=['model', 'system_message', 'n_shot', 'prompt'])]\n",
" predictions_df.to_csv('../data/cookies_llm_eval_predictions.csv', index=False)"
]
},
{
"cell_type": "code",
"execution_count": 179,
"metadata": {},
"outputs": [],
"source": [
"predictions_df = predictions_df[~predictions_df.duplicated(subset=['model', 'system_message', 'n_shot', 'prompt'])]"
]
},
{
"cell_type": "code",
"execution_count": 180,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(400, 12)"
]
},
"execution_count": 180,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"predictions_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 143,
"metadata": {},
"outputs": [],
"source": [
"# import numpy as np\n",
"\n",
"# ids = predictions_df['id'].isna()\n",
"# # apply pd.Series to predictions column for rows where id is not null and change system_message {0,1} to {1,2}\n",
"# new_df_right = predictions_df.loc[ids, 'prediction'].apply(pd.Series)\n",
"# new_df_right['prediction'] = new_df_right['choices'].apply(lambda x: x[0]['message']['content']).drop(columns=['choices'])\n",
"# new_df_left = predictions_df.loc[ids, ['model', 'system_message', 'n_shot', 'prompt', 'completion']].replace({0:1, 1:2})\n",
"# new_df = pd.concat([new_df_left, new_df_right], axis=1)\n",
"\n",
"# predictions_df.columns = ['model', 'system_message', 'n_shot', 'prompt', 'completion', 'id', 'object', 'created', 'openai_model', 'choices', 'usage', 'prediction']\n",
"# new_df.columns = ['model', 'system_message', 'n_shot', 'prompt', 'completion', 'id', 'object', 'created', 'openai_model', 'choices', 'usage', 'prediction']\n",
"# predictions_df.loc[ids] = new_df"
]
},
{
"cell_type": "code",
"execution_count": 155,
"metadata": {},
"outputs": [],
"source": [
"# for col in ['model','system_message','n_shot']:\n",
"# print(predictions_df[col].value_counts())"
]
},
{
"cell_type": "code",
"execution_count": 84,
"metadata": {},
"outputs": [],
"source": [
"# import numpy as np\n",
"\n",
"# # create a copy of predictions_df to manipulate\n",
"# new_predictions_df = predictions_df\n",
"\n",
"# # replace names with 1 or 2\n",
"# def replace_sys_mes_name(x):\n",
"# if x == \"../prompts/gpt4-system-message.txt\":\n",
"# return \"1\"\n",
"# elif x == \"../prompts/gpt4-system-message2.txt\":\n",
"# return \"2\"\n",
"# else:\n",
"# return x\n",
"# new_predictions_df['system_message'] = new_predictions_df['system_message'].apply(lambda x: replace_sys_mes_name(x))\n",
"# # replace None with 0\n",
"# new_predictions_df['n_shot'] = new_predictions_df['n_shot'].apply(lambda x: 0 if x == None or np.nan else x)\n",
"\n",
"# # break up prediction column into sub columns by each of json keys\n",
"# new_predictions_df = pd.concat([new_predictions_df, new_predictions_df['prediction'].apply(pd.Series)], axis=1)"
]
},
{
"cell_type": "code",
"execution_count": 168,
"metadata": {},
"outputs": [],
"source": [
"# predictions_df.drop(columns=['num_correct'], inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 181,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>model</th>\n",
" <th>system_message</th>\n",
" <th>n_shot</th>\n",
" <th>prompt</th>\n",
" <th>completion</th>\n",
" <th>id</th>\n",
" <th>object</th>\n",
" <th>created</th>\n",
" <th>openai_model</th>\n",
" <th>choices</th>\n",
" <th>usage</th>\n",
" <th>prediction</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>gpt-4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>co-2MFE5QVF,Chill Medicated - Watermelon - Syr...</td>\n",
" <td>Chill Medicated,Edible,Beverage,nan,nan</td>\n",
" <td>chatcmpl-7VlTkjAqXNRWfltMPpr5v37uBJIsg</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687805e+09</td>\n",
" <td>gpt-4-0314</td>\n",
" <td>[<OpenAIObject at 0x7fcf7fde94e0> JSON: {\\n \"...</td>\n",
" <td>{\\n \"prompt_tokens\": 54,\\n \"completion_token...</td>\n",
" <td>Hello! It looks like you mentioned a product: ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>gpt-4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>bl-111630024545,Feelz - Space Cowboy 3.5g,nan,...</td>\n",
" <td>Feelz,Flower,Bud,Space Cowboy,3.5</td>\n",
" <td>chatcmpl-7VlTtGF3RGsngfKB1BXufxoTixX2v</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687805e+09</td>\n",
" <td>gpt-4-0314</td>\n",
" <td>[<OpenAIObject at 0x7fcf7f49d2b0> JSON: {\\n \"...</td>\n",
" <td>{\\n \"prompt_tokens\": 51,\\n \"completion_token...</td>\n",
" <td>Hello! It seems like you are referring to a pr...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>gpt-4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>fl-8voAjt83sD,Champelli | Xclusivo 3.5g | Eigh...</td>\n",
" <td>Champelli,Flower,Bud,Xclusivo,3.5</td>\n",
" <td>chatcmpl-7VlU80b0m00VaiGymtj9dbqOggTgR</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687805e+09</td>\n",
" <td>gpt-4-0314</td>\n",
" <td>[<OpenAIObject at 0x7fcf7e306890> JSON: {\\n \"...</td>\n",
" <td>{\\n \"prompt_tokens\": 71,\\n \"completion_token...</td>\n",
" <td>Hello! It seems like you're interested in the ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>gpt-4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>bl-073133213364,CAM - Mellowz #7 7g,nan,FLOWER...</td>\n",
" <td>CAM,Flower,Bud,Mellowz #7,7</td>\n",
" <td>chatcmpl-7VlUHqbsG2kpFHDxAWfsryh6pHmC9</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687805e+09</td>\n",
" <td>gpt-4-0314</td>\n",
" <td>[<OpenAIObject at 0x7fcf7e33d940> JSON: {\\n \"...</td>\n",
" <td>{\\n \"prompt_tokens\": 49,\\n \"completion_token...</td>\n",
" <td>It seems like you are looking for information ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>gpt-4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>fl-fwJQL2AWnS,Backpack Boyz | Bubblegum Gelato...</td>\n",
" <td>Backpack Boyz,Edible,CBD Tincture/Caps/etc,nan...</td>\n",
" <td>chatcmpl-7VlUYvcad2wahIMHavhDEkYrgvjpw</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687805e+09</td>\n",
" <td>gpt-4-0314</td>\n",
" <td>[<OpenAIObject at 0x7fcf7e306980> JSON: {\\n \"...</td>\n",
" <td>{\\n \"prompt_tokens\": 59,\\n \"completion_token...</td>\n",
" <td>Hello! It seems like you are looking for infor...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>...</th>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>395</th>\n",
" <td>gpt-3.5-turbo</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>co-76GP441T,Minntz - Emerald Cut - Indoor - Jo...</td>\n",
" <td>Minntz,Preroll,Joint,Emerald Cut,1</td>\n",
" <td>chatcmpl-7VrjRMvs2l8EJd4PVecpSRPCvV9Hk</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687829e+09</td>\n",
" <td>gpt-3.5-turbo-0301</td>\n",
" <td>[{'index': 0, 'message': {'role': 'assistant',...</td>\n",
" <td>{'prompt_tokens': 125, 'completion_tokens': 23...</td>\n",
" <td>Minntz,Joint,Indoor,Emerald Cut,1g,co-76GP441T.</td>\n",
" </tr>\n",
" <tr>\n",
" <th>396</th>\n",
" <td>gpt-3.5-turbo</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>co-5RAWYHYQ,The Growers Circle - Double Down -...</td>\n",
" <td>The Growers Circle,Flower,Bud,Double Down,3.5</td>\n",
" <td>chatcmpl-7VrjT3wfVoLtq3G6xksfVtLz4FloJ</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687829e+09</td>\n",
" <td>gpt-3.5-turbo-0301</td>\n",
" <td>[{'index': 0, 'message': {'role': 'assistant',...</td>\n",
" <td>{'prompt_tokens': 123, 'completion_tokens': 22...</td>\n",
" <td>The Growers Circle,Double Down,Indoor,3.5g,5RA...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>397</th>\n",
" <td>gpt-3.5-turbo</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>md-1195389,Blue Dream Roll Your Own Sugar Shak...</td>\n",
" <td>Pacific Stone,Flower,Bud,nan,14</td>\n",
" <td>chatcmpl-7VrjVafi1eGBXYfgmGBN0H3b0FzYO</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687829e+09</td>\n",
" <td>gpt-3.5-turbo-0301</td>\n",
" <td>[{'index': 0, 'message': {'role': 'assistant',...</td>\n",
" <td>{'prompt_tokens': 119, 'completion_tokens': 20...</td>\n",
" <td>Pacific Stone,Sugar Shake,Blue Dream,Roll Your...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>398</th>\n",
" <td>gpt-3.5-turbo</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>co-847ZXF37,The Grower Circle - Zoo Dawg x Cos...</td>\n",
" <td>The Growers Circle,Preroll,Joint,Zoo Dawg x Co...</td>\n",
" <td>chatcmpl-7VrjWQpcRxJTdr3f4BUd7totDZpdF</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687829e+09</td>\n",
" <td>gpt-3.5-turbo-0301</td>\n",
" <td>[{'index': 0, 'message': {'role': 'assistant',...</td>\n",
" <td>{'prompt_tokens': 133, 'completion_tokens': 32...</td>\n",
" <td>Multi Joint,Zoo Dawg x Cosa Nostra,The Grower ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>399</th>\n",
" <td>gpt-3.5-turbo</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>co-8EMW15ZM,Flight Bites - S'mores - Gummy - 1...</td>\n",
" <td>Flight Bites,Edible,Gummies,nan,nan</td>\n",
" <td>chatcmpl-7VrjXiUHiyUyH7udPXIjANVmAUrra</td>\n",
" <td>chat.completion</td>\n",
" <td>1.687829e+09</td>\n",
" <td>gpt-3.5-turbo-0301</td>\n",
" <td>[{'index': 0, 'message': {'role': 'assistant',...</td>\n",
" <td>{'prompt_tokens': 129, 'completion_tokens': 21...</td>\n",
" <td>Flight Bites,Gummy,S'mores,10 count,100mg CO₂ ...</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>400 rows × 12 columns</p>\n",
"</div>"
],
"text/plain": [
" model system_message n_shot \\\n",
"0 gpt-4 1 0 \n",
"1 gpt-4 1 0 \n",
"2 gpt-4 1 0 \n",
"3 gpt-4 1 0 \n",
"4 gpt-4 1 0 \n",
".. ... ... ... \n",
"395 gpt-3.5-turbo 2 1 \n",
"396 gpt-3.5-turbo 2 1 \n",
"397 gpt-3.5-turbo 2 1 \n",
"398 gpt-3.5-turbo 2 1 \n",
"399 gpt-3.5-turbo 2 1 \n",
"\n",
" prompt \\\n",
"0 co-2MFE5QVF,Chill Medicated - Watermelon - Syr... \n",
"1 bl-111630024545,Feelz - Space Cowboy 3.5g,nan,... \n",
"2 fl-8voAjt83sD,Champelli | Xclusivo 3.5g | Eigh... \n",
"3 bl-073133213364,CAM - Mellowz #7 7g,nan,FLOWER... \n",
"4 fl-fwJQL2AWnS,Backpack Boyz | Bubblegum Gelato... \n",
".. ... \n",
"395 co-76GP441T,Minntz - Emerald Cut - Indoor - Jo... \n",
"396 co-5RAWYHYQ,The Growers Circle - Double Down -... \n",
"397 md-1195389,Blue Dream Roll Your Own Sugar Shak... \n",
"398 co-847ZXF37,The Grower Circle - Zoo Dawg x Cos... \n",
"399 co-8EMW15ZM,Flight Bites - S'mores - Gummy - 1... \n",
"\n",
" completion \\\n",
"0 Chill Medicated,Edible,Beverage,nan,nan \n",
"1 Feelz,Flower,Bud,Space Cowboy,3.5 \n",
"2 Champelli,Flower,Bud,Xclusivo,3.5 \n",
"3 CAM,Flower,Bud,Mellowz #7,7 \n",
"4 Backpack Boyz,Edible,CBD Tincture/Caps/etc,nan... \n",
".. ... \n",
"395 Minntz,Preroll,Joint,Emerald Cut,1 \n",
"396 The Growers Circle,Flower,Bud,Double Down,3.5 \n",
"397 Pacific Stone,Flower,Bud,nan,14 \n",
"398 The Growers Circle,Preroll,Joint,Zoo Dawg x Co... \n",
"399 Flight Bites,Edible,Gummies,nan,nan \n",
"\n",
" id object created \\\n",
"0 chatcmpl-7VlTkjAqXNRWfltMPpr5v37uBJIsg chat.completion 1.687805e+09 \n",
"1 chatcmpl-7VlTtGF3RGsngfKB1BXufxoTixX2v chat.completion 1.687805e+09 \n",
"2 chatcmpl-7VlU80b0m00VaiGymtj9dbqOggTgR chat.completion 1.687805e+09 \n",
"3 chatcmpl-7VlUHqbsG2kpFHDxAWfsryh6pHmC9 chat.completion 1.687805e+09 \n",
"4 chatcmpl-7VlUYvcad2wahIMHavhDEkYrgvjpw chat.completion 1.687805e+09 \n",
".. ... ... ... \n",
"395 chatcmpl-7VrjRMvs2l8EJd4PVecpSRPCvV9Hk chat.completion 1.687829e+09 \n",
"396 chatcmpl-7VrjT3wfVoLtq3G6xksfVtLz4FloJ chat.completion 1.687829e+09 \n",
"397 chatcmpl-7VrjVafi1eGBXYfgmGBN0H3b0FzYO chat.completion 1.687829e+09 \n",
"398 chatcmpl-7VrjWQpcRxJTdr3f4BUd7totDZpdF chat.completion 1.687829e+09 \n",
"399 chatcmpl-7VrjXiUHiyUyH7udPXIjANVmAUrra chat.completion 1.687829e+09 \n",
"\n",
" openai_model choices \\\n",
"0 gpt-4-0314 [<OpenAIObject at 0x7fcf7fde94e0> JSON: {\\n \"... \n",
"1 gpt-4-0314 [<OpenAIObject at 0x7fcf7f49d2b0> JSON: {\\n \"... \n",
"2 gpt-4-0314 [<OpenAIObject at 0x7fcf7e306890> JSON: {\\n \"... \n",
"3 gpt-4-0314 [<OpenAIObject at 0x7fcf7e33d940> JSON: {\\n \"... \n",
"4 gpt-4-0314 [<OpenAIObject at 0x7fcf7e306980> JSON: {\\n \"... \n",
".. ... ... \n",
"395 gpt-3.5-turbo-0301 [{'index': 0, 'message': {'role': 'assistant',... \n",
"396 gpt-3.5-turbo-0301 [{'index': 0, 'message': {'role': 'assistant',... \n",
"397 gpt-3.5-turbo-0301 [{'index': 0, 'message': {'role': 'assistant',... \n",
"398 gpt-3.5-turbo-0301 [{'index': 0, 'message': {'role': 'assistant',... \n",
"399 gpt-3.5-turbo-0301 [{'index': 0, 'message': {'role': 'assistant',... \n",
"\n",
" usage \\\n",
"0 {\\n \"prompt_tokens\": 54,\\n \"completion_token... \n",
"1 {\\n \"prompt_tokens\": 51,\\n \"completion_token... \n",
"2 {\\n \"prompt_tokens\": 71,\\n \"completion_token... \n",
"3 {\\n \"prompt_tokens\": 49,\\n \"completion_token... \n",
"4 {\\n \"prompt_tokens\": 59,\\n \"completion_token... \n",
".. ... \n",
"395 {'prompt_tokens': 125, 'completion_tokens': 23... \n",
"396 {'prompt_tokens': 123, 'completion_tokens': 22... \n",
"397 {'prompt_tokens': 119, 'completion_tokens': 20... \n",
"398 {'prompt_tokens': 133, 'completion_tokens': 32... \n",
"399 {'prompt_tokens': 129, 'completion_tokens': 21... \n",
"\n",
" prediction \n",
"0 Hello! It looks like you mentioned a product: ... \n",
"1 Hello! It seems like you are referring to a pr... \n",
"2 Hello! It seems like you're interested in the ... \n",
"3 It seems like you are looking for information ... \n",
"4 Hello! It seems like you are looking for infor... \n",
".. ... \n",
"395 Minntz,Joint,Indoor,Emerald Cut,1g,co-76GP441T. \n",
"396 The Growers Circle,Double Down,Indoor,3.5g,5RA... \n",
"397 Pacific Stone,Sugar Shake,Blue Dream,Roll Your... \n",
"398 Multi Joint,Zoo Dawg x Cosa Nostra,The Grower ... \n",
"399 Flight Bites,Gummy,S'mores,10 count,100mg CO₂ ... \n",
"\n",
"[400 rows x 12 columns]"
]
},
"execution_count": 181,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"predictions_df"
]
},
{
"cell_type": "code",
"execution_count": 182,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"669"
]
},
"execution_count": 182,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from util import compare_completion_and_prediction\n",
"\n",
"# Function that uses compare_completion_and_prediction to return num_correct and return zero if there is an error\n",
"def get_num_correct(completion, prediction):\n",
" try:\n",
" return compare_completion_and_prediction(completion, prediction)['num_correct']\n",
" except:\n",
" return 0 # this will be the case when format is incorrect\n",
" \n",
"# Apply get_num_correct function to predictions_df dataframe\n",
"predictions_df['num_correct'] = predictions_df.apply(lambda row: get_num_correct(row['completion'], row['prediction']), axis=1)\n",
"predictions_df['num_correct'].sum() # out of 1000 possible correct predictions (20 samples * 5 cols per sample) * (2 system messages * 2 models * 5 n_shot values)"
]
},
{
"cell_type": "code",
"execution_count": 187,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"model system_message n_shot\n",
"gpt-3.5-turbo 1 0 0.00\n",
" 1 0.00\n",
" 2 0 0.00\n",
"gpt-4 1 0 0.00\n",
" 1 0.00\n",
" 2 0 0.00\n",
"gpt-3.5-turbo 1 2 0.24\n",
" 2 1 0.24\n",
" 2 0.27\n",
" 3 0.36\n",
" 1 3 0.40\n",
" 5 0.44\n",
"gpt-4 2 2 0.45\n",
" 1 2 0.45\n",
" 2 1 0.47\n",
"gpt-3.5-turbo 2 5 0.56\n",
"gpt-4 1 3 0.62\n",
" 2 3 0.67\n",
" 5 0.73\n",
" 1 5 0.79\n",
"Name: num_correct, dtype: float64"
]
},
"execution_count": 187,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"predictions_df.groupby(['model', 'system_message', 'n_shot'])['num_correct'].sum().sort_values() / 100 # out of 100 possible correct predictions (20 samples * 5 cols per sample)"
]
},
{
"cell_type": "code",
"execution_count": 184,
"metadata": {},
"outputs": [],
"source": [
"new_predictions_df.to_csv('../data/cookies_llm_eval_proc_preds.csv', index=False)"
]
},
{
"cell_type": "code",
"execution_count": 76,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<style>\n",
" table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
" .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
" .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
" </style>\n",
"<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>usage/completion_tokens</td><td>▆▆▁▁▁▁▁▁▁▁█▄▁▁▁▁▁▁▁▃▁▁▁▆▂▆▃▅▄▅▆▄▃▁▁▁▁▁▁▁</td></tr><tr><td>usage/elapsed_time</td><td>▄▆▁▁▁▁▂▁▂▁█▃▁▁▁▂▁▁▂▁▁▁▁▄▂▄▂▃▃▄▅▂▁▁▁▁▂▁▁▁</td></tr><tr><td>usage/prompt_tokens</td><td>▁▁▂▂▄▄▆▅██▁▁▃▃▄▅▅██▁▁▃▃▁▁▁▁▁▁▂▁▂▁▄▄▆▆██▁</td></tr><tr><td>usage/total_tokens</td><td>▄▄▂▂▃▃▅▅█▇▆▃▂▂▄▅▅▇▇▂▁▃▂▄▂▄▂▄▃▄▄▃▂▄▃▅▆██▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>usage/completion_tokens</td><td>62</td></tr><tr><td>usage/elapsed_time</td><td>2.40086</td></tr><tr><td>usage/prompt_tokens</td><td>54</td></tr><tr><td>usage/total_tokens</td><td>116</td></tr></table><br/></div></div>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
" View run <strong style=\"color:#cdcd00\">rose-puddle-7</strong> at: <a href='https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval/runs/rbtf91s6' target=\"_blank\">https://wandb.ai/kaleidoscope-data/cookies_llm_experimental_eval/runs/rbtf91s6</a><br/>Synced 6 W&B file(s), 422 media file(s), 422 artifact file(s) and 0 other file(s)"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Find logs at: <code>./wandb/run-20230626_114056-rbtf91s6/logs</code>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"autolog.disable()"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "kd-llm-dc",
"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": "3.10.11"
},
"orig_nbformat": 4
},
"nbformat": 4,
"nbformat_minor": 2
}
|