File size: 99,840 Bytes
b585c7f |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 |
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# h2oGPT API call example\n",
"\n",
"Documentation: https://github.com/h2oai/h2ogpt/blob/main/docs/README_CLIENT.md\n",
"\n",
"Good summary of many of the parameters can be found in the [`grclient.py`](https://github.com/h2oai/h2ogpt/blob/main/gradio_utils/grclient.py) \n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"One can interact with Gradio Client by using either native client or h2oGPT wrapper: \n",
"\n",
"- Using Gradio \\'s native client:\n",
"\n",
" ```python\n",
" from gradio_client import Client\n",
" import ast\n",
" \n",
" HOST_URL = \"http://localhost:7860\"\n",
" client = Client(HOST_URL)\n",
" \n",
" # string of dict for input\n",
" kwargs = dict(instruction_nochat='Who are you?')\n",
" res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
" \n",
" # string of dict for output\n",
" response = ast.literal_eval(res)['response']\n",
" print(response)\n",
" ```\n",
"\n",
"- Using [h2oGPT wrapper for Gradio Native Client](https://github.com/h2oai/h2ogpt/blob/main/docs/README_CLIENT.md#h2ogpt-gradio-wrapper)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Loaded h2oGPT details\n"
]
}
],
"source": [
"from gradio_client import Client\n",
"import ast\n",
"from pprint import pprint\n",
"import json\n",
"from tqdm import tqdm\n",
"from enum import Enum\n",
"\n",
"class LangChainAction(Enum):\n",
" \"\"\"LangChain action\"\"\"\n",
" QUERY = \"Query\"\n",
" SUMMARIZE_MAP = \"Summarize\"\n",
" \n",
"\n",
"with open('../tokens/h2oGPT_details.txt') as f:\n",
" gpt_details = json.load(f)\n",
" print(\"Loaded h2oGPT details\")\n",
"\n",
"# HOST_URL = \"http://localhost:7860\"\n",
"HOST_URL = gpt_details[\"gpt_host_url\"]\n",
"H2OGPT_KEY = gpt_details[\"h2ogpt_key\"]\n",
"LANGCHAIN_MODE = langchain_mode = 'UserData4'\n",
"\n",
"client = Client(HOST_URL)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Utility functions"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"import shutil\n",
"import uuid\n",
"import requests\n",
"from requests.exceptions import HTTPError\n",
"import contextlib\n",
"\n",
"\n",
"def print_full_model_response(response):\n",
" '''\n",
" Helper function to print full response from the h2oGPT call, including all parameters.\n",
" Important keys/parameters:\n",
" - `base_model` - model that used to answer the API call\n",
" - `extra_dict` - model parameters that were used to answer the API call\n",
" - `prompt` - actual prompt sent to LLM\n",
" - `where_from` - how hosted model is running: vLLM , tensor, ....\n",
" '''\n",
" print(\"Model Response with Parameters:\\n\")\n",
" save_dict = ast.literal_eval(res)['save_dict']\n",
" # Remove key from extra_dict\n",
" save_dict.pop('h2ogpt_key', None)\n",
" pprint(save_dict)\n",
" print(\"\\n\")\n",
" try:\n",
" sources = ast.literal_eval(response)['sources']\n",
" print(\"Sources:\\n\")\n",
" pprint(sources)\n",
" print(\"\\n\")\n",
" except:\n",
" print(\"No sources\\n\")\n",
"\n",
"\n",
"def makedirs(path, exist_ok=True, tmp_ok=False, use_base=False):\n",
" \"\"\"\n",
" Avoid some inefficiency in os.makedirs()\n",
" :param path:\n",
" :param exist_ok:\n",
" :param tmp_ok: use /tmp if can't write locally\n",
" :param use_base:\n",
" :return:\n",
" \"\"\"\n",
" if path is None:\n",
" return path\n",
" # if base path set, make relative to that, unless user_path absolute path\n",
" if use_base:\n",
" if os.path.normpath(path) == os.path.normpath(os.path.abspath(path)):\n",
" pass\n",
" else:\n",
" if os.getenv('H2OGPT_BASE_PATH') is not None:\n",
" base_dir = os.path.normpath(os.getenv('H2OGPT_BASE_PATH'))\n",
" path = os.path.normpath(path)\n",
" if not path.startswith(base_dir):\n",
" path = os.path.join(os.getenv('H2OGPT_BASE_PATH', ''), path)\n",
" path = os.path.normpath(path)\n",
"\n",
" if os.path.isdir(path) and os.path.exists(path):\n",
" assert exist_ok, \"Path already exists\"\n",
" return path\n",
" try:\n",
" os.makedirs(path, exist_ok=exist_ok)\n",
" return path\n",
" except FileExistsError:\n",
" # e.g. soft link\n",
" return path\n",
" except PermissionError:\n",
" if tmp_ok:\n",
" path0 = path\n",
" path = os.path.join('/tmp/', path)\n",
" print(\"Permission denied to %s, using %s instead\" % (path0, path), flush=True)\n",
" os.makedirs(path, exist_ok=exist_ok)\n",
" return path\n",
" else:\n",
" raise\n",
"\n",
" \n",
"def shutil_rmtree(*args, **kwargs):\n",
" return shutil.rmtree(*args, **kwargs)\n",
"\n",
"\n",
"def remove(path: str):\n",
" try:\n",
" if path is not None and os.path.exists(path):\n",
" if os.path.isdir(path):\n",
" shutil_rmtree(path, ignore_errors=True)\n",
" else:\n",
" with contextlib.suppress(FileNotFoundError):\n",
" os.remove(path)\n",
" except:\n",
" pass\n",
"\n",
"\n",
"def atomic_move_simple(src, dst):\n",
" try:\n",
" shutil.move(src, dst)\n",
" except (shutil.Error, FileExistsError):\n",
" pass\n",
" remove(src)\n",
"\n",
"\n",
"def download_simple(url, dest=None, overwrite=False, verbose=False):\n",
" if dest is None:\n",
" dest = os.path.basename(url)\n",
" base_path = os.path.dirname(dest)\n",
" if base_path: # else local path\n",
" base_path = makedirs(base_path, exist_ok=True, tmp_ok=True, use_base=True)\n",
" dest = os.path.join(base_path, os.path.basename(dest))\n",
"\n",
" if os.path.isfile(dest):\n",
" if not overwrite:\n",
" print(\"Already have %s from url %s, delete file if invalid\" % (dest, str(url)), flush=True)\n",
" return dest\n",
" else:\n",
" remove(dest)\n",
"\n",
" if verbose:\n",
" print(\"BEGIN get url %s\" % str(url), flush=True)\n",
" if url.startswith(\"file://\"):\n",
" from requests_file import FileAdapter\n",
" s = requests.Session()\n",
" s.mount('file://', FileAdapter())\n",
" url_data = s.get(url, stream=True)\n",
" else:\n",
" url_data = requests.get(url, stream=True)\n",
" if verbose:\n",
" print(\"GOT url %s\" % str(url), flush=True)\n",
"\n",
" if url_data.status_code != requests.codes.ok:\n",
" msg = \"Cannot get url %s, code: %s, reason: %s\" % (\n",
" str(url),\n",
" str(url_data.status_code),\n",
" str(url_data.reason),\n",
" )\n",
" raise requests.exceptions.RequestException(msg)\n",
" url_data.raw.decode_content = True\n",
"\n",
" uuid_tmp = str(uuid.uuid4())[:6]\n",
" dest_tmp = dest + \"_dl_\" + uuid_tmp + \".tmp\"\n",
" with open(dest_tmp, \"wb\") as f:\n",
" shutil.copyfileobj(url_data.raw, f)\n",
" atomic_move_simple(dest_tmp, dest)\n",
" if verbose:\n",
" print(\"DONE url %s\" % str(url), flush=True)\n",
" return dest"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Hello World example"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response:\n",
"\n",
"(\" Hello! My name is LLaMA, I'm a large language model trained by a team of \"\n",
" 'researcher at Meta AI. My primary function is to understand and respond to '\n",
" 'human input in a helpful and engaging manner. I can answer questions, '\n",
" 'provide information, and even generate creative content such as stories or '\n",
" 'dialogue. Is there anything specific you would like to know or talk about?')\n"
]
}
],
"source": [
"# string of dict for input\n",
"kwargs = dict(instruction_nochat='Who are you?',\n",
" h2ogpt_key=H2OGPT_KEY)\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"\n",
"# string of dict for output\n",
"response = ast.literal_eval(res)['response']\n",
"print(\"Model Response:\\n\")\n",
"pprint(response)"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response with Parameters:\n",
"\n",
"{'base_model': 'h2oai/h2ogpt-4096-llama2-70b-chat',\n",
" 'error': '',\n",
" 'extra_dict': {'frequency_penalty': 0,\n",
" 'inference_server': 'vllm:192.176.243.12:5000',\n",
" 'max_tokens': 1024,\n",
" 'n': 1,\n",
" 'ntokens': None,\n",
" 'num_prompt_tokens': 13,\n",
" 'presence_penalty': 0.6,\n",
" 't_generate': 4.012332916259766,\n",
" 'temperature': 0,\n",
" 'tokens_persecond': None,\n",
" 'top_p': 1,\n",
" 'username': 'NO_REQUEST'},\n",
" 'output': \" Hello! My name is LLaMA, I'm a large language model trained by a \"\n",
" 'team of researcher at Meta AI. My primary function is to '\n",
" 'understand and respond to human input in a helpful and engaging '\n",
" 'manner. I can answer questions, provide information, and even '\n",
" 'generate creative content such as stories or dialogue. Is there '\n",
" 'anything specific you would like to know or talk about?',\n",
" 'prompt': '<s>[INST] Who are you? [/INST]',\n",
" 'save_dir': 'saveall_docs',\n",
" 'sources': [],\n",
" 'valid_key': True,\n",
" 'where_from': 'vllm',\n",
" 'which_api': 'str_api'}\n",
"\n",
"\n",
"Sources:\n",
"\n",
"[]\n",
"\n",
"\n"
]
}
],
"source": [
"print_full_model_response(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Setting `temperature` parameter requires setting `do_sample` to `True`. For best reproducibility, set `do_sample` to `False`.\n",
"\n",
"```python"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response:\n",
"\n",
"(\" Hello! I'm LLaMA, an AI assistant developed by Meta AI that can understand \"\n",
" \"and respond to human input in a conversational manner. I'm trained on a \"\n",
" 'massive dataset of text from the internet and can generate human-like '\n",
" 'responses to a wide range of topics and questions. I can be used to create '\n",
" 'chatbots, virtual assistants, and other applications that require natural '\n",
" 'language understanding and generation capabilities.')\n"
]
}
],
"source": [
"# string of dict for input\n",
"kwargs = dict(instruction_nochat='Who are you?',\n",
" seed=123,\n",
" temperature=0.5,\n",
" do_sample=True,\n",
" h2ogpt_key=H2OGPT_KEY)\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"\n",
"# string of dict for output\n",
"response = ast.literal_eval(res)['response']\n",
"print(\"Model Response:\\n\")\n",
"pprint(response)"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response with Parameters:\n",
"\n",
"{'base_model': 'h2oai/h2ogpt-4096-llama2-70b-chat',\n",
" 'error': '',\n",
" 'extra_dict': {'frequency_penalty': 0,\n",
" 'inference_server': 'vllm:192.176.243.12:5000',\n",
" 'max_tokens': 1024,\n",
" 'n': 1,\n",
" 'ntokens': None,\n",
" 'num_prompt_tokens': 13,\n",
" 'presence_penalty': 0.6,\n",
" 't_generate': 3.7804932594299316,\n",
" 'temperature': 0.5,\n",
" 'tokens_persecond': None,\n",
" 'top_p': 0.75,\n",
" 'username': 'NO_REQUEST'},\n",
" 'output': \" Hello! I'm LLaMA, an AI assistant developed by Meta AI that can \"\n",
" 'understand and respond to human input in a conversational manner. '\n",
" \"I'm trained on a massive dataset of text from the internet and can \"\n",
" 'generate human-like responses to a wide range of topics and '\n",
" 'questions. I can be used to create chatbots, virtual assistants, '\n",
" 'and other applications that require natural language understanding '\n",
" 'and generation capabilities.',\n",
" 'prompt': '<s>[INST] Who are you? [/INST]',\n",
" 'save_dir': 'saveall_docs',\n",
" 'sources': [],\n",
" 'valid_key': True,\n",
" 'where_from': 'vllm',\n",
" 'which_api': 'str_api'}\n",
"\n",
"\n",
"Sources:\n",
"\n",
"[]\n",
"\n",
"\n"
]
}
],
"source": [
"print_full_model_response(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Example of Context only call with parameters\n",
"\n",
"Good summary of many of the parameters can be found in the [`grclient.py`](https://github.com/h2oai/h2ogpt/blob/main/gradio_utils/grclient.py) \n",
"\n",
"In the below example, we will set LLM model to use as well as some parameters."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response:\n",
"\n",
"(\" Hello! My name is LLaMA, I'm a large language model trained by a team of \"\n",
" 'researcher at Meta AI. My primary function is to assist with tasks such as '\n",
" 'answering questions, providing information, and generating text. I am '\n",
" 'capable of understanding and responding to human input in a conversational '\n",
" 'manner. I am here to help and provide information to the best of my ability. '\n",
" 'Is there something specific you would like to know or discuss?')\n"
]
}
],
"source": [
"# string of dict for input\n",
"kwargs = dict(instruction_nochat='Who are you?',\n",
" visible_models=['h2oai/h2ogpt-4096-llama2-13b-chat'],\n",
" langchain_mode='LLM',\n",
" max_new_tokens=512,\n",
" max_time=360,\n",
" repetition_penalty=1.07,\n",
" do_sample=True,\n",
" temperature=0.1,\n",
" top_p=0.75,\n",
" penalty_alpha=0,\n",
" h2ogpt_key=H2OGPT_KEY)\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"\n",
"# string of dict for output\n",
"response = ast.literal_eval(res)['response']\n",
"print(\"Model Response:\\n\")\n",
"pprint(response)"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response with Parameters:\n",
"\n",
"{'base_model': 'h2oai/h2ogpt-4096-llama2-13b-chat',\n",
" 'error': '',\n",
" 'extra_dict': {'frequency_penalty': 0,\n",
" 'inference_server': 'vllm:192.176.243.12:5001',\n",
" 'max_tokens': 512,\n",
" 'n': 1,\n",
" 'ntokens': None,\n",
" 'num_prompt_tokens': 13,\n",
" 'presence_penalty': 0.6,\n",
" 't_generate': 2.1190145015716553,\n",
" 'temperature': 0.1,\n",
" 'tokens_persecond': None,\n",
" 'top_p': 0.75,\n",
" 'username': 'NO_REQUEST'},\n",
" 'output': \" Hello! My name is LLaMA, I'm a large language model trained by a \"\n",
" 'team of researcher at Meta AI. My primary function is to assist '\n",
" 'with tasks such as answering questions, providing information, and '\n",
" 'generating text. I am capable of understanding and responding to '\n",
" 'human input in a conversational manner. I am here to help and '\n",
" 'provide information to the best of my ability. Is there something '\n",
" 'specific you would like to know or discuss?',\n",
" 'prompt': '<s>[INST] Who are you? [/INST]',\n",
" 'save_dir': 'saveall_docs',\n",
" 'sources': [],\n",
" 'valid_key': True,\n",
" 'where_from': 'vllm',\n",
" 'which_api': 'str_api'}\n",
"\n",
"\n",
"Sources:\n",
"\n",
"[]\n",
"\n",
"\n"
]
}
],
"source": [
"print_full_model_response(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Summarize Document with mode \"Summarize\"\n",
"\n",
"This approach is useful for the following scenarios:\n",
"- Summarize a given document\n",
"- Ask question about given document. \n",
"\n",
"This is different from asking question (searching) full collection of documents"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 1 - create shared Collection and upload documents\n",
"\n",
"Currently there is no way to authenticate with Gradio Client, therefore we will use shared collection. \n",
"\n",
"The additional examples of Client use can be found in the `test_client_chat_stream_langchain_steps3` function located in the `test_client_calls.py` file. \n",
"\n",
"**Create Shared folder**:"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"user_path = 'user_path'\n",
"new_langchain_mode_text = '%s, %s, %s' % (langchain_mode, 'shared', user_path)\n",
"res = client.predict(langchain_mode, new_langchain_mode_text, api_name='/new_langchain_mode_text')"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"({'__type__': 'update',\n",
" 'choices': [['UserData', 'UserData'],\n",
" ['MyData', 'MyData'],\n",
" ['LLM', 'LLM'],\n",
" ['UserData4', 'UserData4']],\n",
" 'value': 'UserData4'},\n",
" '',\n",
" '/var/folders/_z/jf3ghwdx1kg905xm5p1nktlh0000gp/T/gradio/tmpplv8021u.json')\n"
]
}
],
"source": [
"pprint(res)"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [],
"source": [
"text = \"Yufuu is a wonderful place and you should really visit because there is lots of sun.\"\n",
"loaders = tuple([None, None, None, None])\n",
"res = client.predict(text, langchain_mode, True, 512, True,\n",
" *loaders,\n",
" H2OGPT_KEY,\n",
" api_name='/add_text')"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(None,\n",
" 'UserData4',\n",
" ' <html>\\n'\n",
" ' <body>\\n'\n",
" ' <p>\\n'\n",
" ' Sources: <br>\\n'\n",
" ' </p>\\n'\n",
" ' <div style=\"overflow-y: auto;height:400px\">\\n'\n",
" ' <table>\\n'\n",
" '<thead>\\n'\n",
" '<tr><th style=\"text-align: right;\"> '\n",
" 'index</th><th>source '\n",
" '</th><th>head </th></tr>\\n'\n",
" '</thead>\\n'\n",
" '<tbody>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 1</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_paste/_37aa0924-8.txt\" target=\"_blank\" rel=\"noopener '\n",
" 'noreferrer\">user_paste/_37aa0924-8.txt</a></font></td><td>Yufuu is a '\n",
" 'wonderful place and you should really v</td></tr>\\n'\n",
" '</tbody>\\n'\n",
" '</table>\\n'\n",
" ' </div>\\n'\n",
" ' </body>\\n'\n",
" ' </html>\\n'\n",
" ' ',\n",
" '',\n",
" '_37aa0924-8.txt')\n"
]
}
],
"source": [
"pprint(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Add document to collection via URL"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"url = \"https://www.africau.edu/images/default/sample.pdf\"\n",
"res = client.predict(url,\n",
" langchain_mode, True, 512, True,\n",
" *loaders,\n",
" H2OGPT_KEY,\n",
" api_name='/add_url')"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(None,\n",
" 'UserData4',\n",
" ' <html>\\n'\n",
" ' <body>\\n'\n",
" ' <p>\\n'\n",
" ' Sources: <br>\\n'\n",
" ' </p>\\n'\n",
" ' <div style=\"overflow-y: auto;height:400px\">\\n'\n",
" ' <table>\\n'\n",
" '<thead>\\n'\n",
" '<tr><th style=\"text-align: right;\"> '\n",
" 'index</th><th>source '\n",
" '</th><th>head </th></tr>\\n'\n",
" '</thead>\\n'\n",
" '<tbody>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 1</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_paste/_37aa0924-8.txt\" target=\"_blank\" rel=\"noopener '\n",
" 'noreferrer\">user_paste/_37aa0924-8.txt</a></font> '\n",
" '</td><td>Yufuu is a wonderful place and you should really v</td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 2</td><td><font size=\"2\"><a '\n",
" 'href=\"https://www.africau.edu/images/default/sample.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">https://www.africau.edu/images/default/sample.pdf</a></font></td><td>Simple '\n",
" 'PDF File 2\\n'\n",
" '...continued from page 1. Yet '\n",
" 'mo </td></tr>\\n'\n",
" '</tbody>\\n'\n",
" '</table>\\n'\n",
" ' </div>\\n'\n",
" ' </body>\\n'\n",
" ' </html>\\n'\n",
" ' ',\n",
" '',\n",
" 'sample.pdf')\n"
]
}
],
"source": [
"pprint(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Download file and add to the new collection"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"url = \"https://www.nyserda.ny.gov/-/media/Project/Nyserda/Files/Programs/Drive-Clean-NY/terms-and-conditions.pdf\"\n",
"test_file1 = os.path.join('/tmp/', 'terms-and-conditions.pdf')\n",
"download_simple(url, dest=test_file1)\n",
"\n",
"# upload file(s). Can be list or single file\n",
"# test_file_server - location of the uploaded file on the Gradio server\n",
"test_file_local, test_file_server = client.predict(test_file1, api_name='/upload_api')"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Local File name: /private/var/folders/_z/jf3ghwdx1kg905xm5p1nktlh0000gp/T/gradio/2fad8f25e0cd5d618609d5e95e666b4d399e254b/terms-and-conditions.pdf\n",
"Remote (Gradio Server) File name: /tmp/gradio/55e65c1a447610b8b4ee99717922af03099f9821/terms-and-conditions.pdf\n"
]
}
],
"source": [
"print(\"Local File name:\", test_file_local)\n",
"print(\"Remote (Gradio Server) File name:\", test_file_server)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Add remote file to h2oPT collection"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [],
"source": [
"chunk = True\n",
"chunk_size = 512\n",
"h2ogpt_key = H2OGPT_KEY\n",
"res = client.predict(test_file_server,\n",
" langchain_mode, chunk, chunk_size, True,\n",
" None, None, None, None,\n",
" h2ogpt_key,\n",
" api_name='/add_file_api')"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(None,\n",
" 'UserData4',\n",
" ' <html>\\n'\n",
" ' <body>\\n'\n",
" ' <p>\\n'\n",
" ' Sources: <br>\\n'\n",
" ' </p>\\n'\n",
" ' <div style=\"overflow-y: auto;height:400px\">\\n'\n",
" ' <table>\\n'\n",
" '<thead>\\n'\n",
" '<tr><th style=\"text-align: right;\"> '\n",
" 'index</th><th>source '\n",
" '</th><th>head </th></tr>\\n'\n",
" '</thead>\\n'\n",
" '<tbody>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 1</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_paste/_37aa0924-8.txt\" target=\"_blank\" rel=\"noopener '\n",
" 'noreferrer\">user_paste/_37aa0924-8.txt</a></font> '\n",
" '</td><td>Yufuu is a wonderful place and you should really v</td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 2</td><td><font size=\"2\"><a '\n",
" 'href=\"https://www.africau.edu/images/default/sample.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">https://www.africau.edu/images/default/sample.pdf</a></font></td><td>Simple '\n",
" 'PDF File 2\\n'\n",
" '...continued from page 1. Yet '\n",
" 'mo </td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 3</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_path/terms-and-conditions.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">user_path/terms-and-conditions.pdf</a></font> '\n",
" '</td><td>August 2017\\n'\n",
" '11 I agree to reimburse the dealer '\n",
" 'the </td></tr>\\n'\n",
" '</tbody>\\n'\n",
" '</table>\\n'\n",
" ' </div>\\n'\n",
" ' </body>\\n'\n",
" ' </html>\\n'\n",
" ' ',\n",
" '',\n",
" 'terms-and-conditions.pdf')\n"
]
}
],
"source": [
"pprint(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Add one more file:\n",
"- Upload to Gradio Server\n",
"- Add to Collection"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"url = \"https://cleanvehiclerebate.org/sites/default/files/docs/nav/transportation/cvrp/documents/CVRP-Implementation-Manual.pdf\"\n",
"test_file1 = os.path.join('/tmp/', 'CVRP-Implementation-Manual.pdf')\n",
"download_simple(url, dest=test_file1)\n",
"\n",
"# upload file(s). Can be list or single file\n",
"# test_file_server - location of the uploaded file on the Gradio server\n",
"test_file_local, test_file_server = client.predict(test_file1, api_name='/upload_api')"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [],
"source": [
"chunk = True\n",
"chunk_size = 512\n",
"embed = True\n",
"h2ogpt_key = H2OGPT_KEY\n",
"loaders = tuple([None, None, None, None])\n",
"doc_options = tuple([langchain_mode, chunk, chunk_size, embed])\n",
"\n",
"res = client.predict(\n",
" test_file_server, *doc_options, *loaders, h2ogpt_key, api_name=\"/add_file_api\"\n",
" )"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(None,\n",
" 'UserData4',\n",
" ' <html>\\n'\n",
" ' <body>\\n'\n",
" ' <p>\\n'\n",
" ' Sources: <br>\\n'\n",
" ' </p>\\n'\n",
" ' <div style=\"overflow-y: auto;height:400px\">\\n'\n",
" ' <table>\\n'\n",
" '<thead>\\n'\n",
" '<tr><th style=\"text-align: right;\"> '\n",
" 'index</th><th>source '\n",
" '</th><th>head </th></tr>\\n'\n",
" '</thead>\\n'\n",
" '<tbody>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 1</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_paste/_37aa0924-8.txt\" target=\"_blank\" rel=\"noopener '\n",
" 'noreferrer\">user_paste/_37aa0924-8.txt</a></font> '\n",
" '</td><td>Yufuu is a wonderful place and you should really v</td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 2</td><td><font size=\"2\"><a '\n",
" 'href=\"https://www.africau.edu/images/default/sample.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">https://www.africau.edu/images/default/sample.pdf</a></font></td><td>Simple '\n",
" 'PDF File 2\\n'\n",
" '...continued from page 1. Yet '\n",
" 'mo </td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 3</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_path/terms-and-conditions.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">user_path/terms-and-conditions.pdf</a></font> '\n",
" '</td><td>August 2017\\n'\n",
" '11 I agree to reimburse the dealer '\n",
" 'the </td></tr>\\n'\n",
" '<tr><td style=\"text-align: right;\"> 4</td><td><font size=\"2\"><a '\n",
" 'href=\"file/user_path/CVRP-Implementation-Manual.pdf\" target=\"_blank\" '\n",
" 'rel=\"noopener '\n",
" 'noreferrer\">user_path/CVRP-Implementation-Manual.pdf</a></font> '\n",
" '</td><td>This page intentionally blank. </td></tr>\\n'\n",
" '</tbody>\\n'\n",
" '</table>\\n'\n",
" ' </div>\\n'\n",
" ' </body>\\n'\n",
" ' </html>\\n'\n",
" ' ',\n",
" '',\n",
" 'CVRP-Implementation-Manual.pdf')\n"
]
}
],
"source": [
"pprint(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 2 - retrieve full path to the document already uploaded to h2oGPT\n",
"\n",
"In the below example, we get full path to all documents loaded into \"MyTest\" collection"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['https://www.africau.edu/images/default/sample.pdf',\n",
" 'user_paste/_37aa0924-8.txt',\n",
" 'user_path/CVRP-Implementation-Manual.pdf',\n",
" 'user_path/terms-and-conditions.pdf']\n"
]
}
],
"source": [
"sources = ast.literal_eval(client.predict(langchain_mode, api_name='/get_sources_api'))\n",
"pprint(sources[:10])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 3: Ask questions about the document\n",
"\n",
"PArameters for the LLM input:\n",
"- `pre_prompt_summary` - append to the beginning to the LLM input\n",
"- Document content is sent in between `pre_prompt_summary` and `post_prompt_summary`\n",
"- `prompt_summary` - append to the end of the LLM input"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Summarize single document"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here is a summary of the text in 5 bullet points:\n",
"\n",
"• The Charge NY Drive Clean Rebate Program offers rebates to residents, businesses, fleets, and government entities.\n",
"• The vehicle purchaser must be a New York State resident or business/fleet registered/licensed to do business in New York State.\n",
"• The vehicle purchaser must agree to register/lease the vehicle for at least 36 months in New York State.\n",
"• The vehicle purchaser must agree to participate in online surveys and research efforts and never modify the vehicle's emission control system or engine.\n",
"• The vehicle purchaser must provide accurate information and have the legal authority to commit to the program's obligations.\n"
]
}
],
"source": [
"instruction = None\n",
"document_choice = \"user_path/terms-and-conditions.pdf\"\n",
"\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response with Parameters:\n",
"\n",
"{'base_model': 'h2oai/h2ogpt-4096-llama2-70b-chat',\n",
" 'error': '',\n",
" 'extra_dict': {'add_search_to_context': False,\n",
" 'chat_conversation': [],\n",
" 'context': '',\n",
" 'do_sample': False,\n",
" 'document_choice': 'user_path/terms-and-conditions.pdf',\n",
" 'document_subset': 'Relevant',\n",
" 'early_stopping': False,\n",
" 'iinput': '',\n",
" 'inference_server': 'vllm:192.176.243.12:5000',\n",
" 'instruction': '',\n",
" 'langchain_action': 'Summarize',\n",
" 'langchain_agents': [],\n",
" 'langchain_mode': 'UserData4',\n",
" 'max_new_tokens': 256,\n",
" 'max_time': 360,\n",
" 'min_new_tokens': 0,\n",
" 'ntokens': None,\n",
" 'num_beams': 1,\n",
" 'num_prompt_tokens': 322,\n",
" 'num_return_sequences': 1,\n",
" 'penalty_alpha': 0.0,\n",
" 'prompt_type': 'llama2',\n",
" 'repetition_penalty': 1.07,\n",
" 't_generate': 21.372483015060425,\n",
" 'temperature': 0.1,\n",
" 'tokens_persecond': None,\n",
" 'top_k': 40,\n",
" 'top_p': 0.75,\n",
" 'username': 'NO_REQUEST'},\n",
" 'output': ' Sure! Here is a summary of the text in 5 bullet points:\\n'\n",
" '\\n'\n",
" '• The Charge NY Drive Clean Rebate Program offers rebates to '\n",
" 'residents, businesses, fleets, and government entities.\\n'\n",
" '• The vehicle purchaser must be a New York State resident or '\n",
" 'business/fleet registered/licensed to do business in New York '\n",
" 'State.\\n'\n",
" '• The vehicle purchaser must agree to register/lease the vehicle '\n",
" 'for at least 36 months in New York State.\\n'\n",
" '• The vehicle purchaser must agree to participate in online '\n",
" \"surveys and research efforts and never modify the vehicle's \"\n",
" 'emission control system or engine.\\n'\n",
" '• The vehicle purchaser must provide accurate information and have '\n",
" \"the legal authority to commit to the program's obligations.\",\n",
" 'prompt': '<s>[INST] In order to write a concise single-paragraph or bulleted '\n",
" 'list summary, pay attention to the following text\\n'\n",
" ':\\n'\n",
" '\"\"\"\\n'\n",
" ' Sure! Here is a summary of the text in 5 bullet points:\\n'\n",
" '\\n'\n",
" '• The Charge NY Drive Clean Rebate Program offers rebates to '\n",
" 'residents, businesses, fleets, and government entities that '\n",
" 'purchase or lease eligible vehicles.\\n'\n",
" '• To be eligible, the vehicle purchaser must be a New York State '\n",
" 'resident, government entity, or business/fleet registered/licensed '\n",
" 'to do business in New York State and intends to domicile the '\n",
" 'vehicle in New York State.\\n'\n",
" '• The vehicle purchaser must agree to register/lease the vehicle '\n",
" 'for at least 36 months in New York State, maintain vehicle '\n",
" \"insurance, and allow NYSERDA to verify the vehicle's VIN and \"\n",
" 'registration.\\n'\n",
" '• The vehicle purchaser must also agree to participate in online '\n",
" \"surveys and research efforts, never modify the vehicle's emission \"\n",
" 'control system or engine, and indemnify NYSERDA and the State of '\n",
" 'New York from any liabilities.\\n'\n",
" '• If the vehicle purchaser provides false or inaccurate '\n",
" 'information, they must reimburse the dealer the full value of the '\n",
" 'rebate, and they must have the legal authority to commit to the '\n",
" 'obligations outlined in the program.\\n'\n",
" '\"\"\"\\n'\n",
" 'Using only the text above, write a condensed and concise summary '\n",
" 'of key results as 5 bullet points:\\n'\n",
" ' [/INST]',\n",
" 'save_dir': 'saveall_docs',\n",
" 'sources': [{'content': 'August 2017\\n'\n",
" 'Charge NY Drive Clean Rebate Program\\n'\n",
" 'Vehicle Purchaser Terms and Conditions\\n'\n",
" 'A Vehicle Purchaser is an individual, business, '\n",
" 'fleet, or government entity that purchases or leases '\n",
" 'a vehicle\\n'\n",
" 'that is eligible for a rebate from the Charge NY '\n",
" 'Drive Clean Rebate Program. A Vehicle Purchaser must '\n",
" 'be\\n'\n",
" 'a resident of New York State (if an individual), be '\n",
" 'a New York State government entity or municipality, '\n",
" 'or\\n'\n",
" 'be registered/licensed to do business in New York '\n",
" 'State and must affirm that it intends to domicile '\n",
" 'the\\n'\n",
" 'vehicle in New York State (if a business, fleet, or '\n",
" 'government entity).\\n'\n",
" 'General Terms and Conditions for Vehicle '\n",
" 'Purchasers:\\n'\n",
" 'I hereby acknowledge that I have read and agree to '\n",
" 'meet and follow the requirements and '\n",
" 'responsibilities\\n'\n",
" 'for Vehicle Purchaser participation as set forth '\n",
" 'below.\\n'\n",
" '1\\n'\n",
" 'I certify that I am a New York State Resident, '\n",
" 'government entity or an entity registered/licensed '\n",
" 'to\\n'\n",
" 'do business in New York State.\\n'\n",
" '2\\n'\n",
" 'I certify that the Dealer has explained to me the '\n",
" 'value of the Charge NY Drive Clean Rebate for my\\n'\n",
" 'vehicle purchase and has clearly shown me that the '\n",
" 'full amount of this rebate has been taken off of '\n",
" 'the\\n'\n",
" 'purchase or lease price of the vehicle. I agree to '\n",
" 'allow the Dealer to receive the rebate on my '\n",
" 'behalf.\\n'\n",
" '3\\n'\n",
" 'If I am an individual, I agree to register the '\n",
" 'vehicle with the New York State Department of Motor\\n'\n",
" 'Vehicles with an address located within New York '\n",
" 'State for at least thirty-six (36) months from the\\n'\n",
" 'date of purchase. If I am a Vehicle Purchaser other '\n",
" 'than an individual, I agree to domicile the vehicle\\n'\n",
" 'within New York State for at least thirty-six (36) '\n",
" 'months from the date of purchase. If I leased the\\n'\n",
" 'vehicle, I agree that my original lease term is at '\n",
" 'least thirty-six (36) months.\\n'\n",
" '4\\n'\n",
" 'I agree to allow NYSERDA or its designee to verify '\n",
" 'the vehicle identification number (VIN) and\\n'\n",
" 'registration with the DMV.\\n'\n",
" '5\\n'\n",
" 'I agree to maintain vehicle insurance as required by '\n",
" 'New York State law.\\n'\n",
" '6\\n'\n",
" 'I agree to allow NYSERDA to share my address, '\n",
" 'contact information, and vehicle model purchased\\n'\n",
" 'with the electric distribution utility serving the '\n",
" 'primary location in New York State where the '\n",
" 'vehicle\\n'\n",
" 'will be domiciled for the purpose of informing its '\n",
" 'system planning efforts. I understand that this '\n",
" 'utility\\n'\n",
" 'may send me information about programs that it '\n",
" 'offers to customers that are designed specifically '\n",
" 'for\\n'\n",
" 'plug-in electric vehicle owners.\\n'\n",
" '7\\n'\n",
" 'I agree to never modify the vehicle’s emission '\n",
" 'control system, engine, engine hardware, software\\n'\n",
" 'calibrations, or electric drive system.\\n'\n",
" '8\\n'\n",
" 'I agree to participate in online surveys and other '\n",
" 'research efforts that support Program goals.\\n'\n",
" '9\\n'\n",
" 'I acknowledge that neither NYSERDA, nor any of its '\n",
" 'consultants, is responsible for assuring that the\\n'\n",
" 'vehicle is proper for the Vehicle Purchaser or '\n",
" 'complies with any particular laws, codes, or '\n",
" 'industry\\n'\n",
" 'standards. I acknowledge that NYSERDA has made no '\n",
" 'representations of any kind regarding the\\n'\n",
" 'results to be achieved by the Program.\\n'\n",
" '10 I shall protect, indemnify and hold harmless '\n",
" 'NYSERDA and the State of New York from and against\\n'\n",
" 'all liabilities, losses, claims, damages, judgments, '\n",
" 'penalties, causes of action, costs and expenses\\n'\n",
" \"(including, without limitation, attorneys' fees and \"\n",
" 'expenses) imposed upon or incurred by or asserted\\n'\n",
" 'against NYSERDA or the State of New York resulting '\n",
" 'from, arising out of or relating to Vehicle\\n'\n",
" 'Purchaser’s participation in the Program including, '\n",
" 'without limitation, Vehicle Purchaser’s purchase\\n'\n",
" 'or lease of vehicles in association therewith;\\n'\n",
" '\\n'\n",
" 'August 2017\\n'\n",
" '11 I agree to reimburse the dealer the full value of '\n",
" 'the rebate if it is discovered that I provided false '\n",
" 'or\\n'\n",
" 'inaccurate information that results in the rebate '\n",
" 'application being denied; and\\n'\n",
" '12 I certify that I have the legal authority to '\n",
" 'commit the Vehicle Purchaser to the obligations '\n",
" 'herein.\\n'\n",
" 'If the Vehicle Purchaser is an individual, fill out '\n",
" 'this section:\\n'\n",
" 'Name of Vehicle Purchaser: '\n",
" '_________________________________\\n'\n",
" 'Signature of Vehicle Purchaser: '\n",
" '_______________________________\\n'\n",
" 'Email of Vehicle Purchaser: '\n",
" '__________________________________\\n'\n",
" 'Date: ___________________________\\n'\n",
" 'Scan a copy of the Vehicle Purchaser’s New York '\n",
" 'State Driver’s License and include it in the box '\n",
" 'below\\n'\n",
" 'or upload a copy as a separate document in Step 6 of '\n",
" 'the online rebate application:\\n'\n",
" 'If the Vehicle Purchaser is a non-individual (fleet, '\n",
" 'business, or government entity), fill out this '\n",
" 'section:\\n'\n",
" 'Legal Business Name (Government Name): '\n",
" '_______________________________\\n'\n",
" 'Employer Identification Number: '\n",
" '_______________________________\\n'\n",
" 'New York State address where the vehicle will be '\n",
" 'domiciled:\\n'\n",
" '_______________________________\\n'\n",
" '_______________________________\\n'\n",
" '_______________________________\\n'\n",
" 'Name and Title of Authorized Representative: '\n",
" '_______________________________\\n'\n",
" 'Signature of Authorized Representative: '\n",
" '____________________________________\\n'\n",
" 'Email of Authorized Representative: '\n",
" '_______________________________________\\n'\n",
" 'Date: ___________________________',\n",
" 'orig_index': 0,\n",
" 'score': 0,\n",
" 'source': 'user_path/terms-and-conditions.pdf'}],\n",
" 'valid_key': True,\n",
" 'where_from': 'run_qa_db',\n",
" 'which_api': 'str_api'}\n",
"\n",
"\n",
"Sources:\n",
"\n",
"[{'content': 'August 2017\\n'\n",
" 'Charge NY Drive Clean Rebate Program\\n'\n",
" 'Vehicle Purchaser Terms and Conditions\\n'\n",
" 'A Vehicle Purchaser is an individual, business, fleet, or '\n",
" 'government entity that purchases or leases a vehicle\\n'\n",
" 'that is eligible for a rebate from the Charge NY Drive Clean '\n",
" 'Rebate Program. A Vehicle Purchaser must be\\n'\n",
" 'a resident of New York State (if an individual), be a New York '\n",
" 'State government entity or municipality, or\\n'\n",
" 'be registered/licensed to do business in New York State and must '\n",
" 'affirm that it intends to domicile the\\n'\n",
" 'vehicle in New York State (if a business, fleet, or government '\n",
" 'entity).\\n'\n",
" 'General Terms and Conditions for Vehicle Purchasers:\\n'\n",
" 'I hereby acknowledge that I have read and agree to meet and '\n",
" 'follow the requirements and responsibilities\\n'\n",
" 'for Vehicle Purchaser participation as set forth below.\\n'\n",
" '1\\n'\n",
" 'I certify that I am a New York State Resident, government entity '\n",
" 'or an entity registered/licensed to\\n'\n",
" 'do business in New York State.\\n'\n",
" '2\\n'\n",
" 'I certify that the Dealer has explained to me the value of the '\n",
" 'Charge NY Drive Clean Rebate for my\\n'\n",
" 'vehicle purchase and has clearly shown me that the full amount '\n",
" 'of this rebate has been taken off of the\\n'\n",
" 'purchase or lease price of the vehicle. I agree to allow the '\n",
" 'Dealer to receive the rebate on my behalf.\\n'\n",
" '3\\n'\n",
" 'If I am an individual, I agree to register the vehicle with the '\n",
" 'New York State Department of Motor\\n'\n",
" 'Vehicles with an address located within New York State for at '\n",
" 'least thirty-six (36) months from the\\n'\n",
" 'date of purchase. If I am a Vehicle Purchaser other than an '\n",
" 'individual, I agree to domicile the vehicle\\n'\n",
" 'within New York State for at least thirty-six (36) months from '\n",
" 'the date of purchase. If I leased the\\n'\n",
" 'vehicle, I agree that my original lease term is at least '\n",
" 'thirty-six (36) months.\\n'\n",
" '4\\n'\n",
" 'I agree to allow NYSERDA or its designee to verify the vehicle '\n",
" 'identification number (VIN) and\\n'\n",
" 'registration with the DMV.\\n'\n",
" '5\\n'\n",
" 'I agree to maintain vehicle insurance as required by New York '\n",
" 'State law.\\n'\n",
" '6\\n'\n",
" 'I agree to allow NYSERDA to share my address, contact '\n",
" 'information, and vehicle model purchased\\n'\n",
" 'with the electric distribution utility serving the primary '\n",
" 'location in New York State where the vehicle\\n'\n",
" 'will be domiciled for the purpose of informing its system '\n",
" 'planning efforts. I understand that this utility\\n'\n",
" 'may send me information about programs that it offers to '\n",
" 'customers that are designed specifically for\\n'\n",
" 'plug-in electric vehicle owners.\\n'\n",
" '7\\n'\n",
" 'I agree to never modify the vehicle’s emission control system, '\n",
" 'engine, engine hardware, software\\n'\n",
" 'calibrations, or electric drive system.\\n'\n",
" '8\\n'\n",
" 'I agree to participate in online surveys and other research '\n",
" 'efforts that support Program goals.\\n'\n",
" '9\\n'\n",
" 'I acknowledge that neither NYSERDA, nor any of its consultants, '\n",
" 'is responsible for assuring that the\\n'\n",
" 'vehicle is proper for the Vehicle Purchaser or complies with any '\n",
" 'particular laws, codes, or industry\\n'\n",
" 'standards. I acknowledge that NYSERDA has made no '\n",
" 'representations of any kind regarding the\\n'\n",
" 'results to be achieved by the Program.\\n'\n",
" '10 I shall protect, indemnify and hold harmless NYSERDA and the '\n",
" 'State of New York from and against\\n'\n",
" 'all liabilities, losses, claims, damages, judgments, penalties, '\n",
" 'causes of action, costs and expenses\\n'\n",
" \"(including, without limitation, attorneys' fees and expenses) \"\n",
" 'imposed upon or incurred by or asserted\\n'\n",
" 'against NYSERDA or the State of New York resulting from, arising '\n",
" 'out of or relating to Vehicle\\n'\n",
" 'Purchaser’s participation in the Program including, without '\n",
" 'limitation, Vehicle Purchaser’s purchase\\n'\n",
" 'or lease of vehicles in association therewith;\\n'\n",
" '\\n'\n",
" 'August 2017\\n'\n",
" '11 I agree to reimburse the dealer the full value of the rebate '\n",
" 'if it is discovered that I provided false or\\n'\n",
" 'inaccurate information that results in the rebate application '\n",
" 'being denied; and\\n'\n",
" '12 I certify that I have the legal authority to commit the '\n",
" 'Vehicle Purchaser to the obligations herein.\\n'\n",
" 'If the Vehicle Purchaser is an individual, fill out this '\n",
" 'section:\\n'\n",
" 'Name of Vehicle Purchaser: _________________________________\\n'\n",
" 'Signature of Vehicle Purchaser: _______________________________\\n'\n",
" 'Email of Vehicle Purchaser: __________________________________\\n'\n",
" 'Date: ___________________________\\n'\n",
" 'Scan a copy of the Vehicle Purchaser’s New York State Driver’s '\n",
" 'License and include it in the box below\\n'\n",
" 'or upload a copy as a separate document in Step 6 of the online '\n",
" 'rebate application:\\n'\n",
" 'If the Vehicle Purchaser is a non-individual (fleet, business, '\n",
" 'or government entity), fill out this section:\\n'\n",
" 'Legal Business Name (Government Name): '\n",
" '_______________________________\\n'\n",
" 'Employer Identification Number: _______________________________\\n'\n",
" 'New York State address where the vehicle will be domiciled:\\n'\n",
" '_______________________________\\n'\n",
" '_______________________________\\n'\n",
" '_______________________________\\n'\n",
" 'Name and Title of Authorized Representative: '\n",
" '_______________________________\\n'\n",
" 'Signature of Authorized Representative: '\n",
" '____________________________________\\n'\n",
" 'Email of Authorized Representative: '\n",
" '_______________________________________\\n'\n",
" 'Date: ___________________________',\n",
" 'orig_index': 0,\n",
" 'score': 0,\n",
" 'source': 'user_path/terms-and-conditions.pdf'}]\n",
"\n",
"\n"
]
}
],
"source": [
"print_full_model_response(res)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Additional Single document summary"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here's a summary of the text in 5 bullet points:\n",
"\n",
"• A simple PDF file is being demonstrated.\n",
"• The file contains a lot of text, described as boring.\n",
"• The file is being used for Virtual Mechanics tutorials.\n",
"• The author finds typing the text boring.\n",
"• The author mentions that watching paint dry is even more boring.\n"
]
}
],
"source": [
"instruction = None\n",
"document_choice = \"https://www.africau.edu/images/default/sample.pdf\"\n",
"\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Summarize California EV program"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here is a summary of the key points in 5 bullet points:\n",
"\n",
"• The Clean Vehicle Rebate Project (CVRP) provides rebates for purchasing or leasing eligible zero- and near-zero-emission vehicles.\n",
"• CVRP is administered by the California Air Resources Board (CARB) and aims to encourage the development and deployment of advanced technologies.\n",
"• Funding for the CVRP comes from the Greenhouse Gas Reduction Fund.\n",
"• The program outlines minimum requirements for implementation in the CVRP Terms and Conditions, Guidelines, and Funding Plan.\n",
"• The program benefits disadvantaged communities.\n"
]
}
],
"source": [
"instruction = None\n",
"document_choice = \"user_path/CVRP-Implementation-Manual.pdf\"\n",
"\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Summarize all documents in the Collection\n"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here is a summary of the key points in 5 bullet points:\n",
"\n",
"• The Clean Vehicle Rebate Project (CVRP) provides rebates for purchasing or leasing eligible zero- and near-zero-emission vehicles.\n",
"• CVRP is administered by the California Air Resources Board (CARB) and aims to encourage the development and deployment of advanced technologies that reduce greenhouse gas emissions.\n",
"• Funding for the CVRP comes from the Greenhouse Gas Reduction Fund.\n",
"• The program benefits California citizens by providing immediate air pollution emission reductions.\n",
"• The program promotes the development of cleaner vehicles.\n"
]
}
],
"source": [
"instruction = None\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" #document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Question answering for a single document\n",
"\n",
"We will use summary mode as well, even though we are not summarizing the document. \n",
"This mode will enable us to send full document for question answering task."
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here is a summary of the eligibility criteria for the Charge NY Drive Clean Rebate Program:\n",
"\n",
"• You must be a resident of New York State, a New York State government entity, or registered/licensed to do business in New York State.\n",
"• You must purchase or lease a vehicle that is eligible for a rebate from the Charge NY Drive Clean Rebate Program.\n",
"• You must register the vehicle with the New York State Department of Motor Vehicles with an address located within New York State for at least thirty-six (36) months from the date of purchase.\n",
"• You must allow NYSERDA or its designee to verify the vehicle identification number (VIN) and registration with the DMV.\n",
"• You must maintain vehicle insurance as required by New York State law.\n",
"\n",
"Does this help?\n"
]
}
],
"source": [
"instruction = \"What is the eligibility criteria for the program?\"\n",
"document_choice = \"user_path/terms-and-conditions.pdf\"\n",
"\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"\n",
"# pre_prompt_query = \"\"\"Pay attention and remember the information below, which will help to answer the question or imperative after the context ends.\\n\"\"\"\n",
"# prompt_query = \"\"\"According to only the information in the document sources provided within the context above, \\n\"\"\"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Question answering for all documents in the Collection"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" According to the information provided in the context, the eligibility criteria for the Clean Vehicle Rebate Project (CVRP) includes:\n",
"\n",
"1. Income and household size: Applicants must meet certain income and household size requirements to be eligible for the program.\n",
"2. Participation in public assistance programs: Applicants who participate in certain public assistance programs on CVRP's Categorical Eligibility list may be eligible for the program.\n",
"3. Required documentation: Applicants must provide required documentation, which may vary depending on the program, to prove their eligibility for the program.\n",
"4. Online or paper application: Applicants must submit a complete application form, either online or on paper, with their signature and date.\n",
"5. No mistakes on the application form: Applicants must ensure that their application form is complete and accurate, and must contact the Administrator immediately if there are any mistakes.\n",
"6. Updates to governing documents: Applicants must be aware of updates to CVRP governing documents, which can affect their eligibility for the program.\n",
"\n",
"It is important to note that these are the general el\n"
]
}
],
"source": [
"instruction = \"What is the eligibility criteria for the program?\"\n",
"document_choice = \"user_path/terms-and-conditions.pdf\"\n",
"\n",
"langchain_action = LangChainAction.QUERY.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"#pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"#prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"pre_prompt_summary = None\n",
"prompt_summary = None\n",
"\n",
"pre_prompt_query = \"\"\"Pay attention and remember the information below, which will help to answer the question or imperative after the context ends.\\n\"\"\"\n",
"prompt_query = \"\"\"According to only the information in the document sources provided within the context above, \\n\"\"\"\n",
"#pre_prompt_query = None\n",
"#prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" # document_choice=document_choice,\n",
" max_new_tokens=256,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model Response with Parameters:\n",
"\n",
"{'base_model': 'h2oai/h2ogpt-4096-llama2-70b-chat',\n",
" 'error': '',\n",
" 'extra_dict': {'add_search_to_context': False,\n",
" 'chat_conversation': [],\n",
" 'context': '',\n",
" 'do_sample': False,\n",
" 'document_choice': ['All'],\n",
" 'document_subset': 'Relevant',\n",
" 'early_stopping': False,\n",
" 'iinput': '',\n",
" 'inference_server': 'vllm:192.176.243.12:5000',\n",
" 'instruction': 'What is the eligibility criteria for the '\n",
" 'program?',\n",
" 'langchain_action': 'Query',\n",
" 'langchain_agents': [],\n",
" 'langchain_mode': 'UserData4',\n",
" 'max_new_tokens': 256,\n",
" 'max_time': 360,\n",
" 'min_new_tokens': 0,\n",
" 'ntokens': None,\n",
" 'num_beams': 1,\n",
" 'num_prompt_tokens': 514,\n",
" 'num_return_sequences': 1,\n",
" 'penalty_alpha': 0.0,\n",
" 'prompt_type': 'llama2',\n",
" 'repetition_penalty': 1.07,\n",
" 't_generate': 12.24568796157837,\n",
" 'temperature': 0.1,\n",
" 'tokens_persecond': None,\n",
" 'top_k': 40,\n",
" 'top_p': 0.75,\n",
" 'username': 'NO_REQUEST'},\n",
" 'output': ' According to the information provided in the context, the '\n",
" 'eligibility criteria for the Clean Vehicle Rebate Project (CVRP) '\n",
" 'includes:\\n'\n",
" '\\n'\n",
" '1. Income and household size: Applicants must meet certain income '\n",
" 'and household size requirements to be eligible for the program.\\n'\n",
" '2. Participation in public assistance programs: Applicants who '\n",
" \"participate in certain public assistance programs on CVRP's \"\n",
" 'Categorical Eligibility list may be eligible for the program.\\n'\n",
" '3. Required documentation: Applicants must provide required '\n",
" 'documentation, which may vary depending on the program, to prove '\n",
" 'their eligibility for the program.\\n'\n",
" '4. Online or paper application: Applicants must submit a complete '\n",
" 'application form, either online or on paper, with their signature '\n",
" 'and date.\\n'\n",
" '5. No mistakes on the application form: Applicants must ensure '\n",
" 'that their application form is complete and accurate, and must '\n",
" 'contact the Administrator immediately if there are any mistakes.\\n'\n",
" '6. Updates to governing documents: Applicants must be aware of '\n",
" 'updates to CVRP governing documents, which can affect their '\n",
" 'eligibility for the program.\\n'\n",
" '\\n'\n",
" 'It is important to note that these are the general el',\n",
" 'prompt': '<s>[INST] \\n'\n",
" '\"\"\"\\n'\n",
" 'Pay attention and remember the information below, which will help '\n",
" 'to answer the question or imperative after the context ends.\\n'\n",
" 'on income and household size.\\n'\n",
" 'If an applicant applying for an increased rebate participates in '\n",
" 'one or more of the\\n'\n",
" 'public assistance programs on CVRP’s Categorical Eligibility list, '\n",
" 'they may submit\\n'\n",
" 'documentation confirming their current participation for '\n",
" 'consideration by the\\n'\n",
" 'Administrator in lieu of IRS Form 4506-C. Note that depending on '\n",
" 'the program,\\n'\n",
" 'documentation required may vary.\\n'\n",
" '• Required documentation for public fleet pre-acquisition '\n",
" 'reservations will also\\n'\n",
" 'include the following:\\n'\n",
" '\\n'\n",
" 'document to either obtain money or property from the State or '\n",
" 'avoid paying or\\n'\n",
" 'transmitting money or property to the State. CARB also retains the '\n",
" 'authority to\\n'\n",
" 'prohibit any entity from participating in CVRP due to '\n",
" 'noncompliance with project\\n'\n",
" 'requirements or fraud which includes attempted fraud.\\n'\n",
" 'During the application process, applicants should provide to the '\n",
" 'Administrator all\\n'\n",
" 'information necessary for the assessment of their applications. '\n",
" 'Applicants whose\\n'\n",
" '\\n'\n",
" 'Required documentation will include, at a minimum, the following:\\n'\n",
" '• For online applicants, you will be required to date and type '\n",
" 'your name (which will\\n'\n",
" 'act as your signature) on the submitted application form. This '\n",
" 'signed and dated\\n'\n",
" 'document is required. For applicants who request a paper '\n",
" 'application form, a\\n'\n",
" 'complete application with signature and date. Contact the '\n",
" 'Administrator\\n'\n",
" 'immediately if there is a mistake on your application form. '\n",
" 'Applicants who submit\\n'\n",
" '\\n'\n",
" 'eligibility for the Clean Vehicle Rebate Project (CVRP). These '\n",
" 'governing documents\\n'\n",
" 'are updated several times every year to accommodate operational '\n",
" 'process changes\\n'\n",
" 'and may affect the applicant’s eligibility for the program. The '\n",
" 'next scheduled updates\\n'\n",
" 'to CVRP governing documents can be found in the CVRP FAQs at\\n'\n",
" 'CleanVehicleRebate.org/FAQs under “How often do CVRP program '\n",
" 'requirements\\n'\n",
" 'change?” Note that CVRP reserves the right to update the '\n",
" 'Implementation Manual\\n'\n",
" '\\n'\n",
" 'eligibility for the CVRP rebate program.\\n'\n",
" '\"\"\"\\n'\n",
" 'According to only the information in the document sources provided '\n",
" 'within the context above, \\n'\n",
" 'What is the eligibility criteria for the program? [/INST]',\n",
" 'save_dir': 'saveall_docs',\n",
" 'sources': [{'content': 'on income and household size.\\n'\n",
" 'If an applicant applying for an increased rebate '\n",
" 'participates in one or more of the\\n'\n",
" 'public assistance programs on CVRP’s Categorical '\n",
" 'Eligibility list, they may submit\\n'\n",
" 'documentation confirming their current participation '\n",
" 'for consideration by the\\n'\n",
" 'Administrator in lieu of IRS Form 4506-C. Note that '\n",
" 'depending on the program,\\n'\n",
" 'documentation required may vary.\\n'\n",
" '• Required documentation for public fleet '\n",
" 'pre-acquisition reservations will also\\n'\n",
" 'include the following:',\n",
" 'orig_index': 1,\n",
" 'score': 0.2838561339693884,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'document to either obtain money or property from the '\n",
" 'State or avoid paying or\\n'\n",
" 'transmitting money or property to the State. CARB '\n",
" 'also retains the authority to\\n'\n",
" 'prohibit any entity from participating in CVRP due '\n",
" 'to noncompliance with project\\n'\n",
" 'requirements or fraud which includes attempted '\n",
" 'fraud.\\n'\n",
" 'During the application process, applicants should '\n",
" 'provide to the Administrator all\\n'\n",
" 'information necessary for the assessment of their '\n",
" 'applications. Applicants whose',\n",
" 'orig_index': 3,\n",
" 'score': 0.2903084456920624,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'Required documentation will include, at a minimum, '\n",
" 'the following:\\n'\n",
" '• For online applicants, you will be required to '\n",
" 'date and type your name (which will\\n'\n",
" 'act as your signature) on the submitted application '\n",
" 'form. This signed and dated\\n'\n",
" 'document is required. For applicants who request a '\n",
" 'paper application form, a\\n'\n",
" 'complete application with signature and date. '\n",
" 'Contact the Administrator\\n'\n",
" 'immediately if there is a mistake on your '\n",
" 'application form. Applicants who submit',\n",
" 'orig_index': 4,\n",
" 'score': 0.29075086265597117,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'eligibility for the Clean Vehicle Rebate Project '\n",
" '(CVRP). These governing documents\\n'\n",
" 'are updated several times every year to accommodate '\n",
" 'operational process changes\\n'\n",
" 'and may affect the applicant’s eligibility for the '\n",
" 'program. The next scheduled updates\\n'\n",
" 'to CVRP governing documents can be found in the CVRP '\n",
" 'FAQs at\\n'\n",
" 'CleanVehicleRebate.org/FAQs under “How often do CVRP '\n",
" 'program requirements\\n'\n",
" 'change?” Note that CVRP reserves the right to update '\n",
" 'the Implementation Manual',\n",
" 'orig_index': 2,\n",
" 'score': 0.2900393307209015,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'eligibility for the CVRP rebate program.',\n",
" 'orig_index': 0,\n",
" 'score': 0.21977069973945618,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'}],\n",
" 'valid_key': True,\n",
" 'where_from': 'run_qa_db',\n",
" 'which_api': 'str_api'}\n",
"\n",
"\n",
"Sources:\n",
"\n",
"[{'content': 'on income and household size.\\n'\n",
" 'If an applicant applying for an increased rebate participates in '\n",
" 'one or more of the\\n'\n",
" 'public assistance programs on CVRP’s Categorical Eligibility '\n",
" 'list, they may submit\\n'\n",
" 'documentation confirming their current participation for '\n",
" 'consideration by the\\n'\n",
" 'Administrator in lieu of IRS Form 4506-C. Note that depending on '\n",
" 'the program,\\n'\n",
" 'documentation required may vary.\\n'\n",
" '• Required documentation for public fleet pre-acquisition '\n",
" 'reservations will also\\n'\n",
" 'include the following:',\n",
" 'orig_index': 1,\n",
" 'score': 0.2838561339693884,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'document to either obtain money or property from the State or '\n",
" 'avoid paying or\\n'\n",
" 'transmitting money or property to the State. CARB also retains '\n",
" 'the authority to\\n'\n",
" 'prohibit any entity from participating in CVRP due to '\n",
" 'noncompliance with project\\n'\n",
" 'requirements or fraud which includes attempted fraud.\\n'\n",
" 'During the application process, applicants should provide to the '\n",
" 'Administrator all\\n'\n",
" 'information necessary for the assessment of their applications. '\n",
" 'Applicants whose',\n",
" 'orig_index': 3,\n",
" 'score': 0.2903084456920624,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'Required documentation will include, at a minimum, the '\n",
" 'following:\\n'\n",
" '• For online applicants, you will be required to date and type '\n",
" 'your name (which will\\n'\n",
" 'act as your signature) on the submitted application form. This '\n",
" 'signed and dated\\n'\n",
" 'document is required. For applicants who request a paper '\n",
" 'application form, a\\n'\n",
" 'complete application with signature and date. Contact the '\n",
" 'Administrator\\n'\n",
" 'immediately if there is a mistake on your application form. '\n",
" 'Applicants who submit',\n",
" 'orig_index': 4,\n",
" 'score': 0.29075086265597117,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'eligibility for the Clean Vehicle Rebate Project (CVRP). These '\n",
" 'governing documents\\n'\n",
" 'are updated several times every year to accommodate operational '\n",
" 'process changes\\n'\n",
" 'and may affect the applicant’s eligibility for the program. The '\n",
" 'next scheduled updates\\n'\n",
" 'to CVRP governing documents can be found in the CVRP FAQs at\\n'\n",
" 'CleanVehicleRebate.org/FAQs under “How often do CVRP program '\n",
" 'requirements\\n'\n",
" 'change?” Note that CVRP reserves the right to update the '\n",
" 'Implementation Manual',\n",
" 'orig_index': 2,\n",
" 'score': 0.2900393307209015,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'},\n",
" {'content': 'eligibility for the CVRP rebate program.',\n",
" 'orig_index': 0,\n",
" 'score': 0.21977069973945618,\n",
" 'source': 'user_path/CVRP-Implementation-Manual.pdf'}]\n",
"\n",
"\n"
]
}
],
"source": [
"print_full_model_response(res)"
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" Sure! Here's a summary of the income eligibility criteria for the program based on the provided text:\n",
"\n",
"• The CVRP's income eligibility criteria are based on gross annual household income.\n",
"• The maximum income eligibility levels are ﹩135,000 for single filers, ﹩175,000 for head-of-household filers, and ﹩200,000 for joint filers.\n",
"• Applicants who are claimed as dependents are not eligible for increased rebates regardless of their income.\n",
"• Income verification is completed using IRS Form 1040 and/or other proof of income documentation.\n",
"• The income cap applies to all eligible vehicle types except FCEVs.\n"
]
}
],
"source": [
"instruction = \"What is the income eligibility criteria for the program?\"\n",
"document_choice = \"user_path/CVRP-Implementation-Manual.pdf\"\n",
"langchain_action = LangChainAction.SUMMARIZE_MAP.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"#pre_prompt_summary = None\n",
"#prompt_summary = None\n",
"\n",
"#pre_prompt_query = \"\"\"Pay attention and remember the information below, which will help to answer the question or imperative after the context ends.\\n\"\"\"\n",
"#prompt_query = \"\"\"According to only the information in the document sources provided within the context above, \\n\"\"\"\n",
"pre_prompt_query = None\n",
"prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" document_choice=document_choice,\n",
" max_new_tokens=1026,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Ask Collection and question and get answers for all documents in the collection"
]
},
{
"cell_type": "code",
"execution_count": 47,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" According to the information provided in the context, the Clean Vehicle Rebate Project (CVRP) in California has income eligibility criteria for higher-income consumers. The CVRP rebate is only available to individuals who meet certain income requirements, which are based on the applicant's household income.\n",
"\n",
"The income eligibility criteria for the CVRP rebate are as follows:\n",
"\n",
"* For households with a gross annual income of ﹩150,000 or less, the rebate is available for the full amount of ﹩2,500.\n",
"* For households with a gross annual income between ﹩150,001 and ﹩200,000, the rebate is reduced by 50%.\n",
"* For households with a gross annual income between ﹩200,001 and ﹩250,000, the rebate is reduced by 75%.\n",
"* For households with a gross annual income of ﹩250,001 or more, the rebate is not available.\n",
"\n",
"It's important to note that these income eligibility criteria are subject to change, and the CVRP may have additional requirements or restrictions. It's always best to check the program's website or contact the CVRP directly for the most up-to-date information on income eligibility criteria and other program requirements.\n"
]
}
],
"source": [
"instruction = \"What is the income eligibility criteria for the Clean Vehicle Rebate Project in the state of California?\"\n",
"langchain_action = LangChainAction.QUERY.value\n",
"stream_output = False\n",
"top_k_docs = 5\n",
"\n",
"#pre_prompt_summary = \"\"\"In order to write a concise single-paragraph or bulleted list summary, pay attention to the following text\\n\"\"\"\n",
"#prompt_summary = \"Using only the text above, write a condensed and concise summary of key results as 5 bullet points:\\n\"\n",
"pre_prompt_summary = None\n",
"prompt_summary = None\n",
"\n",
"pre_prompt_query = \"\"\"Pay attention and remember the information below, which will help to answer the question or imperative after the context ends.\\n\"\"\"\n",
"prompt_query = \"\"\"According to only the information in the document sources provided within the context above, \\n\"\"\"\n",
"#pre_prompt_query = None\n",
"#prompt_query = None\n",
"\n",
"kwargs = dict(instruction=instruction,\n",
" langchain_mode=langchain_mode,\n",
" langchain_action=langchain_action, # uses full document, not vectorDB chunks\n",
" top_k_docs=top_k_docs,\n",
" stream_output=stream_output,\n",
" document_subset='Relevant',\n",
" # document_choice=document_choice,\n",
" max_new_tokens=1026,\n",
" max_time=360,\n",
" do_sample=False,\n",
" pre_prompt_query=pre_prompt_query,\n",
" prompt_query=prompt_query,\n",
" pre_prompt_summary=pre_prompt_summary,\n",
" prompt_summary=prompt_summary,\n",
" h2ogpt_key=H2OGPT_KEY\n",
" )\n",
"\n",
"# get result\n",
"res = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
"response = ast.literal_eval(res)['response']\n",
"print(response)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "gen-ai-python310",
"language": "python",
"name": "gen-ai-python310"
},
"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
}
|