glm4 results

logs/glm-4-9b-chat-1m_epoch_1.txt

@@ -150,8 +150,8 @@ step 4: 不是
 step 5: 不是
 (3) GPU = NVIDIA L40. Max memory = 44.309 GB.
 18.473 GB of memory reserved.
-      text label  ... THUDM/glm-4-9b-chat-1m THUDM/glm-4-9b-chat-1m_checkpoint-175
-0  甄加索是自杀吗    不是  ...                  回答：不是                                    不是
+      text  ... THUDM/glm-4-9b-chat-1m_checkpoint-700
+0  甄加索是自杀吗  ...                                    不是
 
-[1 rows x 7 columns]
-{'accuracy': 0.5946666666666667, 'incorrect_ids': [5, 7, 9, 15, 16, 17, 18, 24, 25, 26, 27, 28, 31, 33, 34, 35, 36, 42, 43, 48, 55, 58, 59, 61, 65, 66, 67, 69, 71, 75, 77, 78, 82, 83, 85, 90, 91, 93, 94, 95, 96, 101, 102, 104, 105, 106, 111, 112, 113, 115, 118, 120, 121, 122, 123, 126, 127, 128, 131, 135, 136, 137, 138, 139, 140, 150, 153, 155, 156, 161, 162, 164, 170, 171, 172, 174, 175, 178, 179, 180, 181, 184, 186, 187, 188, 189, 190, 191, 192, 193, 199, 200, 202, 209, 214, 216, 221, 224, 225, 226, 227, 228, 230, 233, 234, 236, 239, 240, 243, 245, 250, 251, 252, 256, 257, 260, 261, 262, 263, 265, 268, 271, 272, 273, 278, 283, 284, 286, 289, 292, 293, 294, 295, 299, 301, 304, 305, 307, 308, 309, 311, 314, 315, 317, 320, 323, 325, 326, 328, 332, 333, 334, 335, 337, 338, 339, 340, 344, 346, 347, 348, 349, 350, 352, 353, 355, 356, 358, 361, 362, 364, 365, 367, 368, 373, 374, 376, 377, 378, 380, 381, 389, 395, 396, 397, 401, 402, 410, 421, 425, 428, 429, 430, 431, 432, 436, 438, 440, 442, 445, 447, 450, 452, 454, 455, 456, 457, 459, 460, 464, 465, 471, 472, 473, 475, 476, 482, 488, 490, 492, 494, 498, 500, 501, 502, 504, 506, 507, 508, 510, 511, 514, 518, 519, 530, 531, 533, 536, 538, 540, 543, 553, 560, 566, 567, 568, 570, 571, 579, 581, 589, 591, 592, 593, 594, 597, 598, 599, 601, 609, 613, 614, 621, 626, 627, 629, 631, 632, 633, 635, 637, 641, 643, 644, 645, 646, 647, 649, 651, 653, 655, 656, 657, 663, 665, 666, 667, 668, 669, 671, 674, 678, 680, 682, 683, 686, 691, 693, 695, 699, 701, 702, 703, 704, 705, 708, 716, 717, 719, 720, 721, 722, 724, 725, 726, 730, 733, 734, 735, 739, 740, 741, 743, 747, 752, 754, 758, 759, 763, 765, 766, 768, 770, 772, 773, 774, 776, 778, 781, 782, 785, 791, 795, 798, 799, 801, 803, 805, 809, 810, 812, 813, 814, 815, 817, 820, 821, 822, 823, 824, 827, 828, 832, 833, 834, 840, 843, 847, 850, 851, 853, 856, 861, 866, 867, 869, 870, 871, 873, 876, 877, 878, 884, 886, 888, 889, 890, 891, 892, 894, 897, 901, 904, 907, 910, 912, 913, 917, 920, 927, 928, 929, 936, 937, 942, 945, 946, 953, 954, 956, 957, 965, 968, 969, 970, 972, 973, 980, 981, 983, 987, 989, 991, 993, 998, 1001, 1003, 1006, 1007, 1009, 1011, 1012, 1014, 1015, 1022, 1023, 1025, 1028, 1029, 1036, 1038, 1039, 1040, 1041, 1043, 1045, 1048, 1050, 1051, 1054, 1056, 1061, 1062, 1063, 1069, 1073, 1074, 1075, 1077, 1078, 1079, 1080, 1081, 1082, 1084, 1087, 1089, 1091, 1096, 1097, 1098, 1099, 1102, 1105, 1106, 1107, 1111, 1114, 1116, 1120, 1121, 1125, 1126, 1128, 1129, 1130, 1133, 1134, 1138, 1139, 1140, 1141, 1143, 1147, 1151, 1153, 1156, 1159, 1161, 1163, 1166, 1168, 1172, 1175, 1177, 1183, 1185, 1192, 1196, 1203, 1207, 1209, 1212, 1219, 1220, 1224, 1228, 1230, 1231, 1232, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1243, 1245, 1246, 1249, 1252, 1254, 1256, 1259, 1264, 1266, 1267, 1270, 1272, 1274, 1276, 1278, 1281, 1282, 1286, 1289, 1292, 1299, 1300, 1305, 1306, 1308, 1312, 1313, 1314, 1317, 1318, 1320, 1321, 1322, 1323, 1328, 1329, 1331, 1334, 1337, 1342, 1345, 1348, 1349, 1353, 1354, 1357, 1358, 1361, 1362, 1363, 1364, 1371, 1372, 1373, 1378, 1379, 1385, 1386, 1387, 1388, 1389, 1390, 1391, 1392, 1395, 1400, 1402, 1406, 1416, 1417, 1420, 1422, 1424, 1425, 1426, 1427, 1428, 1430, 1431, 1432, 1435, 1438, 1439, 1440, 1441, 1443, 1444, 1445, 1450, 1451, 1452, 1453, 1454, 1455, 1459, 1462, 1466, 1467, 1469, 1472, 1478, 1482, 1484, 1485, 1487, 1489, 1490, 1491, 1493, 1495, 1496, 1498, 1503, 1504, 1510, 1512, 1516, 1517, 1518, 1522, 1525, 1526, 1528, 1533, 1536, 1542, 1543, 1544, 1545, 1546, 1547, 1551, 1552, 1556, 1558, 1560, 1561, 1562, 1568, 1569, 1577, 1581, 1585, 1586, 1587, 1588, 1590, 1591, 1593, 1594, 1602, 1603, 1604, 1605, 1606, 1607, 1611, 1612, 1613, 1620, 1621, 1622, 1623, 1624, 1627, 1631, 1632, 1633, 1636, 1637, 1639, 1641, 1643, 1645, 1648, 1652, 1653, 1654, 1655, 1658, 1660, 1668, 1672, 1677, 1682, 1684, 1690, 1691, 1694, 1695, 1699, 1711, 1713, 1716, 1717, 1718, 1723, 1727, 1728, 1735, 1738, 1743, 1745, 1751, 1753, 1754, 1755, 1756, 1760, 1763, 1765, 1766, 1767, 1769, 1770, 1772, 1777, 1778, 1779, 1780, 1785, 1786, 1788, 1791, 1796, 1797, 1799, 1805, 1810, 1812, 1816, 1821, 1822, 1823, 1824, 1834, 1835, 1840, 1841, 1846, 1848, 1849, 1852, 1853, 1855, 1860, 1870, 1872, 1879, 1882, 1884, 1890, 1894, 1896, 1897, 1903, 1909, 1912, 1913, 1916, 1919, 1926, 1928, 1935, 1936, 1938, 1940, 1941, 1944, 1949, 1951, 1953, 1956, 1957, 1958, 1959, 1961, 1963, 1964, 1970, 1973, 1975, 1978, 1981, 1982, 1985, 1986, 1988, 1989, 1990, 1992, 1993, 1995, 1996, 2002, 2003, 2008, 2013, 2014, 2015, 2017, 2019, 2020, 2021, 2024, 2025, 2029, 2034, 2035, 2038, 2042, 2043, 2044, 2046, 2047, 2049, 2052, 2053, 2058, 2059, 2061, 2062, 2064, 2069, 2072, 2073, 2077, 2079, 2081, 2082, 2088, 2093, 2095, 2100, 2107, 2109, 2111, 2112, 2119, 2121, 2126, 2127, 2130, 2131, 2132, 2133, 2139, 2142, 2146, 2149, 2159, 2160, 2161, 2162, 2167, 2168, 2169, 2171, 2173, 2177, 2179, 2180, 2183, 2184, 2185, 2186, 2188, 2191, 2192, 2193, 2194, 2195, 2196, 2197, 2198, 2199, 2200, 2209, 2212, 2213, 2215, 2219, 2222, 2223, 2224, 2226, 2230, 2231, 2233, 2240, 2243, 2245, 2247, 2250, 2251, 2254, 2260, 2261, 2262, 2266, 2267, 2274, 2276, 2278, 2279, 2280, 2281, 2282, 2285, 2290, 2293, 2301, 2309, 2311, 2313, 2317, 2320, 2321, 2322, 2324, 2329, 2332, 2333, 2334, 2335, 2336, 2338, 2339, 2340, 2344, 2348, 2352, 2356, 2359, 2360, 2361, 2362, 2364, 2366, 2367, 2369, 2370, 2376, 2378, 2379, 2383, 2388, 2390, 2395, 2398, 2400, 2404, 2405, 2406, 2409, 2410, 2412, 2419, 2420, 2421, 2423, 2424, 2425, 2429, 2435, 2437, 2440, 2441, 2442, 2444, 2445, 2446, 2448, 2450, 2452, 2464, 2465, 2472, 2475, 2477, 2484, 2488, 2496, 2497, 2499, 2501, 2503, 2511, 2512, 2515, 2516, 2517, 2518, 2521, 2523, 2524, 2525, 2528, 2529, 2530, 2532, 2538, 2539, 2540, 2546, 2550, 2551, 2552, 2556, 2557, 2559, 2571, 2574, 2575, 2577, 2589, 2592, 2593, 2594, 2595, 2597, 2600, 2602, 2604, 2605, 2606, 2607, 2608, 2609, 2610, 2616, 2620, 2622, 2623, 2624, 2625, 2626, 2628, 2629, 2630, 2633, 2635, 2636, 2638, 2642, 2644, 2645, 2647, 2652, 2653, 2655, 2657, 2660, 2663, 2664, 2665, 2667, 2671, 2672, 2675, 2677, 2678, 2685, 2686, 2687, 2694, 2697, 2702, 2703, 2707, 2713, 2714, 2715, 2716, 2717, 2726, 2727, 2730, 2731, 2732, 2733, 2740, 2742, 2744, 2746, 2747, 2749, 2754, 2757, 2760, 2761, 2762, 2764, 2765, 2766, 2767, 2768, 2769, 2771, 2777, 2781, 2783, 2788, 2789, 2791, 2792, 2793, 2795, 2796, 2798, 2801, 2803, 2805, 2806, 2807, 2810, 2813, 2814, 2815, 2816, 2818, 2820, 2821, 2822, 2823, 2824, 2829, 2833, 2837, 2838, 2841, 2843, 2844, 2845, 2846, 2851, 2854, 2857, 2860, 2861, 2865, 2870, 2871, 2876, 2877, 2878, 2880, 2881, 2882, 2884, 2885, 2898, 2899, 2902, 2904, 2905, 2910, 2912, 2915, 2916, 2917, 2920, 2921, 2923, 2928, 2933, 2935, 2937, 2942, 2944, 2946, 2949, 2953, 2954, 2955, 2960, 2965, 2967, 2968, 2972, 2976, 2978, 2979, 2980, 2983, 2985, 2991, 2998]}
+[1 rows x 10 columns]
+{'accuracy': 0.596, 'incorrect_ids': [7, 9, 10, 12, 15, 16, 17, 24, 26, 27, 28, 31, 33, 34, 35, 36, 38, 58, 60, 65, 66, 67, 69, 71, 72, 76, 78, 82, 83, 85, 90, 93, 94, 95, 97, 101, 102, 103, 104, 105, 109, 111, 112, 113, 115, 118, 121, 123, 127, 128, 131, 135, 136, 137, 138, 139, 141, 144, 150, 153, 155, 156, 161, 163, 164, 169, 170, 174, 175, 178, 179, 180, 188, 189, 190, 193, 199, 200, 202, 209, 214, 218, 221, 222, 223, 224, 225, 226, 227, 228, 230, 233, 234, 236, 237, 239, 243, 245, 250, 251, 252, 253, 254, 257, 260, 262, 263, 265, 268, 271, 272, 273, 276, 278, 279, 284, 286, 292, 293, 294, 295, 301, 305, 307, 308, 309, 311, 314, 315, 317, 320, 321, 323, 325, 326, 328, 330, 332, 333, 334, 335, 337, 338, 340, 344, 347, 348, 349, 350, 351, 353, 354, 355, 356, 360, 361, 362, 364, 365, 367, 368, 370, 371, 372, 373, 374, 375, 377, 378, 389, 395, 396, 397, 401, 410, 414, 419, 421, 425, 428, 429, 430, 431, 432, 436, 438, 445, 447, 452, 454, 456, 458, 459, 460, 461, 464, 465, 469, 471, 472, 473, 475, 476, 478, 480, 482, 487, 490, 492, 494, 496, 498, 500, 501, 502, 503, 504, 506, 507, 508, 509, 510, 511, 512, 513, 517, 518, 520, 530, 536, 538, 540, 543, 553, 560, 566, 567, 568, 570, 571, 579, 581, 589, 591, 592, 593, 597, 598, 599, 601, 612, 613, 614, 621, 625, 627, 628, 629, 632, 637, 641, 642, 643, 646, 649, 650, 653, 655, 656, 657, 661, 663, 665, 666, 668, 670, 671, 674, 676, 678, 682, 683, 686, 688, 690, 691, 693, 694, 695, 696, 701, 702, 703, 705, 708, 709, 714, 717, 718, 719, 721, 722, 723, 729, 730, 731, 732, 733, 734, 740, 743, 747, 752, 754, 757, 758, 760, 766, 769, 770, 772, 773, 774, 784, 785, 791, 795, 798, 800, 801, 803, 805, 809, 810, 812, 813, 815, 817, 818, 819, 820, 821, 823, 824, 827, 832, 833, 834, 835, 837, 840, 842, 843, 847, 848, 849, 851, 856, 861, 866, 867, 869, 870, 871, 873, 876, 878, 886, 889, 890, 891, 894, 896, 897, 901, 903, 904, 907, 909, 910, 913, 920, 927, 928, 936, 937, 941, 945, 950, 953, 954, 956, 957, 964, 965, 966, 969, 970, 972, 973, 980, 982, 986, 987, 991, 998, 1001, 1003, 1006, 1011, 1012, 1014, 1020, 1022, 1025, 1028, 1029, 1036, 1038, 1040, 1043, 1046, 1048, 1051, 1053, 1054, 1061, 1062, 1063, 1064, 1069, 1073, 1074, 1075, 1076, 1077, 1078, 1079, 1080, 1081, 1084, 1087, 1089, 1096, 1098, 1102, 1106, 1107, 1111, 1114, 1120, 1121, 1123, 1124, 1125, 1129, 1130, 1133, 1134, 1141, 1143, 1149, 1151, 1153, 1161, 1163, 1164, 1166, 1167, 1168, 1172, 1174, 1175, 1177, 1181, 1183, 1185, 1192, 1196, 1203, 1209, 1212, 1216, 1219, 1220, 1224, 1225, 1227, 1228, 1231, 1232, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1243, 1245, 1246, 1249, 1252, 1254, 1256, 1258, 1259, 1261, 1262, 1264, 1267, 1269, 1270, 1273, 1274, 1289, 1292, 1296, 1299, 1304, 1305, 1306, 1308, 1309, 1312, 1313, 1314, 1316, 1320, 1322, 1326, 1328, 1329, 1331, 1334, 1336, 1337, 1342, 1345, 1348, 1349, 1353, 1354, 1358, 1363, 1367, 1372, 1376, 1379, 1380, 1386, 1387, 1389, 1391, 1392, 1393, 1400, 1406, 1407, 1409, 1413, 1416, 1417, 1418, 1420, 1422, 1424, 1425, 1426, 1427, 1428, 1430, 1432, 1438, 1439, 1440, 1441, 1443, 1444, 1445, 1448, 1450, 1451, 1452, 1453, 1455, 1456, 1457, 1459, 1462, 1463, 1467, 1469, 1472, 1473, 1475, 1476, 1477, 1478, 1482, 1484, 1485, 1487, 1488, 1489, 1490, 1491, 1493, 1494, 1495, 1496, 1500, 1504, 1507, 1510, 1512, 1516, 1517, 1518, 1525, 1526, 1527, 1528, 1533, 1536, 1542, 1543, 1545, 1546, 1547, 1551, 1552, 1558, 1560, 1561, 1562, 1565, 1568, 1569, 1577, 1581, 1585, 1587, 1590, 1591, 1593, 1594, 1596, 1598, 1600, 1602, 1603, 1604, 1605, 1607, 1611, 1612, 1613, 1616, 1617, 1621, 1622, 1623, 1624, 1625, 1627, 1631, 1633, 1636, 1637, 1641, 1645, 1647, 1648, 1650, 1652, 1654, 1655, 1658, 1660, 1662, 1666, 1668, 1669, 1672, 1676, 1682, 1683, 1684, 1690, 1691, 1694, 1695, 1700, 1701, 1709, 1711, 1713, 1716, 1717, 1718, 1723, 1727, 1728, 1734, 1735, 1736, 1738, 1740, 1743, 1751, 1753, 1754, 1755, 1756, 1759, 1767, 1769, 1770, 1772, 1774, 1778, 1779, 1780, 1785, 1786, 1789, 1790, 1792, 1796, 1797, 1799, 1810, 1811, 1812, 1816, 1818, 1823, 1824, 1827, 1830, 1834, 1835, 1840, 1846, 1848, 1849, 1852, 1853, 1856, 1858, 1860, 1862, 1870, 1879, 1880, 1881, 1882, 1884, 1888, 1890, 1894, 1896, 1897, 1903, 1904, 1909, 1913, 1917, 1919, 1920, 1922, 1928, 1930, 1931, 1934, 1935, 1936, 1941, 1949, 1951, 1953, 1955, 1956, 1957, 1959, 1961, 1963, 1964, 1970, 1973, 1978, 1979, 1981, 1982, 1984, 1986, 1988, 1989, 1990, 1992, 1995, 1996, 1997, 2002, 2019, 2020, 2021, 2025, 2028, 2030, 2034, 2035, 2036, 2038, 2042, 2043, 2044, 2046, 2047, 2052, 2053, 2059, 2061, 2062, 2063, 2064, 2066, 2069, 2072, 2075, 2077, 2081, 2084, 2088, 2089, 2092, 2094, 2100, 2102, 2104, 2105, 2106, 2109, 2114, 2115, 2118, 2119, 2120, 2121, 2122, 2126, 2128, 2131, 2132, 2133, 2135, 2138, 2139, 2141, 2142, 2144, 2145, 2147, 2149, 2152, 2159, 2160, 2161, 2162, 2169, 2171, 2173, 2175, 2177, 2179, 2181, 2184, 2185, 2186, 2191, 2192, 2195, 2196, 2198, 2199, 2200, 2203, 2204, 2205, 2209, 2212, 2214, 2219, 2221, 2223, 2224, 2226, 2229, 2230, 2231, 2233, 2234, 2235, 2240, 2242, 2243, 2245, 2246, 2247, 2250, 2251, 2253, 2256, 2260, 2261, 2262, 2267, 2274, 2276, 2278, 2280, 2282, 2285, 2287, 2290, 2293, 2297, 2302, 2309, 2311, 2316, 2320, 2321, 2322, 2324, 2328, 2329, 2333, 2335, 2337, 2339, 2348, 2350, 2351, 2352, 2353, 2359, 2361, 2362, 2363, 2364, 2366, 2367, 2368, 2373, 2376, 2378, 2379, 2388, 2390, 2395, 2396, 2398, 2400, 2404, 2406, 2408, 2409, 2410, 2412, 2413, 2420, 2421, 2423, 2425, 2426, 2429, 2435, 2437, 2440, 2441, 2442, 2444, 2445, 2446, 2448, 2457, 2461, 2463, 2464, 2465, 2469, 2472, 2475, 2479, 2481, 2482, 2488, 2491, 2495, 2496, 2498, 2499, 2511, 2512, 2515, 2516, 2518, 2521, 2523, 2526, 2528, 2529, 2530, 2532, 2538, 2539, 2540, 2541, 2545, 2546, 2547, 2548, 2549, 2551, 2552, 2556, 2559, 2562, 2571, 2574, 2575, 2577, 2586, 2587, 2589, 2592, 2593, 2594, 2595, 2597, 2602, 2606, 2607, 2608, 2613, 2616, 2619, 2620, 2621, 2622, 2623, 2624, 2625, 2626, 2628, 2629, 2633, 2635, 2636, 2638, 2642, 2644, 2645, 2652, 2653, 2654, 2657, 2659, 2663, 2664, 2665, 2666, 2667, 2670, 2671, 2672, 2675, 2676, 2678, 2682, 2686, 2687, 2689, 2691, 2692, 2694, 2697, 2701, 2702, 2703, 2707, 2708, 2713, 2714, 2715, 2716, 2717, 2718, 2720, 2726, 2730, 2731, 2732, 2734, 2735, 2737, 2739, 2740, 2742, 2744, 2746, 2747, 2749, 2751, 2754, 2756, 2760, 2761, 2762, 2764, 2766, 2767, 2769, 2772, 2775, 2778, 2779, 2781, 2788, 2791, 2792, 2793, 2796, 2798, 2801, 2803, 2805, 2807, 2810, 2811, 2814, 2815, 2816, 2820, 2821, 2823, 2824, 2831, 2833, 2837, 2841, 2843, 2844, 2845, 2846, 2851, 2857, 2858, 2860, 2861, 2863, 2864, 2865, 2870, 2871, 2874, 2877, 2878, 2879, 2880, 2882, 2884, 2885, 2892, 2895, 2897, 2898, 2899, 2902, 2904, 2905, 2912, 2916, 2919, 2921, 2923, 2926, 2928, 2933, 2935, 2937, 2942, 2944, 2949, 2960, 2965, 2967, 2968, 2969, 2973, 2975, 2976, 2978, 2979, 2980, 2981, 2982, 2985, 2991, 2998]}

logs/glm-4-9b-chat-1m_epoch_2.txt

@@ -150,8 +150,8 @@ step 4: 不是
 step 5: 不是
 (3) GPU = NVIDIA L40. Max memory = 44.309 GB.
 18.473 GB of memory reserved.
-      text  ... THUDM/glm-4-9b-chat-1m_checkpoint-350
+      text  ... THUDM/glm-4-9b-chat-1m_checkpoint-700
 0  甄加索是自杀吗  ...                                    不是
 
-[1 rows x 8 columns]
-{'accuracy': 0.549, 'incorrect_ids': [4, 7, 9, 10, 12, 17, 24, 27, 28, 31, 34, 43, 52, 55, 56, 58, 59, 60, 61, 65, 66, 67, 70, 71, 76, 77, 78, 83, 84, 99, 101, 104, 105, 107, 108, 112, 115, 117, 118, 120, 121, 123, 124, 127, 131, 135, 136, 138, 139, 143, 150, 152, 153, 155, 160, 164, 165, 169, 173, 174, 175, 178, 179, 180, 184, 185, 187, 188, 189, 190, 193, 200, 202, 203, 204, 209, 214, 215, 216, 218, 222, 223, 224, 226, 227, 228, 232, 233, 236, 237, 240, 243, 245, 248, 250, 251, 252, 253, 257, 258, 261, 262, 263, 268, 269, 271, 273, 275, 278, 284, 286, 289, 293, 294, 299, 303, 304, 305, 308, 312, 314, 315, 317, 321, 328, 330, 333, 334, 335, 338, 339, 340, 341, 344, 347, 348, 349, 350, 351, 352, 354, 355, 356, 359, 362, 364, 367, 368, 371, 372, 373, 374, 376, 378, 381, 386, 387, 389, 396, 402, 414, 416, 419, 423, 425, 428, 429, 430, 431, 436, 438, 440, 442, 445, 447, 450, 452, 454, 456, 458, 459, 460, 461, 463, 464, 471, 472, 473, 476, 479, 480, 481, 483, 488, 490, 493, 494, 496, 498, 500, 501, 502, 503, 506, 507, 508, 509, 511, 513, 514, 516, 517, 518, 531, 536, 538, 543, 552, 553, 560, 566, 568, 571, 576, 579, 581, 585, 589, 591, 592, 593, 594, 596, 598, 601, 612, 613, 614, 621, 625, 628, 629, 632, 636, 643, 644, 646, 647, 655, 663, 665, 666, 667, 668, 669, 671, 673, 674, 675, 676, 678, 680, 682, 683, 689, 690, 691, 694, 695, 701, 702, 705, 706, 708, 709, 711, 712, 717, 718, 719, 720, 721, 722, 729, 730, 733, 734, 735, 740, 741, 743, 747, 752, 754, 755, 757, 758, 759, 760, 763, 765, 766, 769, 770, 771, 772, 773, 774, 775, 777, 778, 782, 784, 785, 788, 790, 791, 794, 795, 798, 799, 801, 802, 805, 812, 813, 814, 817, 818, 820, 821, 822, 823, 824, 826, 827, 828, 831, 832, 833, 834, 837, 840, 841, 843, 844, 845, 847, 848, 850, 856, 857, 861, 863, 866, 869, 870, 871, 873, 876, 878, 882, 884, 888, 889, 890, 891, 892, 893, 894, 895, 896, 897, 899, 901, 904, 906, 907, 908, 910, 914, 917, 920, 926, 927, 929, 930, 932, 933, 937, 939, 940, 941, 942, 944, 945, 946, 949, 953, 954, 956, 959, 961, 962, 964, 969, 970, 971, 972, 973, 980, 982, 983, 984, 986, 988, 991, 993, 995, 998, 1000, 1001, 1003, 1005, 1007, 1009, 1011, 1012, 1014, 1015, 1020, 1022, 1023, 1025, 1028, 1031, 1036, 1038, 1039, 1043, 1044, 1045, 1046, 1051, 1053, 1054, 1056, 1058, 1060, 1061, 1063, 1066, 1069, 1071, 1072, 1073, 1074, 1077, 1078, 1080, 1081, 1083, 1084, 1086, 1087, 1089, 1091, 1093, 1096, 1098, 1099, 1102, 1106, 1107, 1110, 1111, 1113, 1114, 1115, 1116, 1117, 1121, 1123, 1126, 1128, 1129, 1134, 1135, 1136, 1140, 1141, 1143, 1144, 1151, 1153, 1155, 1156, 1159, 1163, 1166, 1168, 1170, 1171, 1172, 1173, 1174, 1177, 1178, 1179, 1182, 1183, 1185, 1186, 1192, 1196, 1200, 1203, 1205, 1206, 1209, 1212, 1213, 1216, 1219, 1220, 1224, 1225, 1227, 1228, 1230, 1231, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1244, 1245, 1249, 1251, 1252, 1256, 1259, 1263, 1264, 1265, 1266, 1267, 1269, 1270, 1271, 1272, 1274, 1278, 1281, 1285, 1286, 1289, 1292, 1293, 1299, 1300, 1301, 1304, 1305, 1308, 1312, 1313, 1314, 1317, 1318, 1319, 1320, 1321, 1322, 1324, 1326, 1329, 1331, 1334, 1336, 1337, 1342, 1344, 1346, 1348, 1349, 1352, 1353, 1357, 1358, 1359, 1360, 1362, 1363, 1364, 1365, 1367, 1368, 1371, 1372, 1374, 1378, 1379, 1380, 1384, 1385, 1386, 1387, 1389, 1390, 1391, 1392, 1394, 1395, 1400, 1405, 1406, 1407, 1409, 1413, 1416, 1419, 1420, 1422, 1424, 1425, 1428, 1430, 1432, 1437, 1438, 1439, 1440, 1444, 1445, 1448, 1450, 1451, 1452, 1453, 1454, 1455, 1457, 1459, 1462, 1466, 1467, 1469, 1473, 1475, 1476, 1477, 1478, 1482, 1487, 1488, 1489, 1490, 1492, 1493, 1495, 1496, 1499, 1500, 1502, 1504, 1506, 1507, 1508, 1510, 1512, 1515, 1516, 1517, 1518, 1519, 1520, 1524, 1525, 1526, 1528, 1532, 1533, 1536, 1537, 1539, 1540, 1542, 1544, 1545, 1546, 1547, 1549, 1554, 1556, 1558, 1559, 1560, 1561, 1562, 1563, 1567, 1568, 1574, 1577, 1578, 1579, 1580, 1581, 1584, 1585, 1589, 1590, 1591, 1593, 1594, 1600, 1601, 1602, 1603, 1604, 1605, 1606, 1607, 1608, 1611, 1616, 1617, 1618, 1619, 1620, 1625, 1627, 1631, 1632, 1633, 1636, 1637, 1638, 1639, 1641, 1645, 1648, 1654, 1655, 1657, 1658, 1659, 1668, 1669, 1671, 1672, 1673, 1674, 1678, 1682, 1684, 1686, 1688, 1690, 1691, 1694, 1695, 1697, 1698, 1700, 1701, 1708, 1710, 1712, 1716, 1718, 1721, 1722, 1726, 1727, 1729, 1730, 1734, 1736, 1737, 1738, 1739, 1740, 1743, 1745, 1747, 1751, 1753, 1754, 1755, 1756, 1759, 1761, 1763, 1764, 1765, 1766, 1767, 1768, 1770, 1771, 1776, 1778, 1779, 1780, 1785, 1786, 1788, 1796, 1798, 1799, 1803, 1804, 1805, 1809, 1810, 1812, 1818, 1824, 1827, 1834, 1835, 1836, 1837, 1839, 1841, 1843, 1845, 1846, 1848, 1849, 1851, 1857, 1859, 1860, 1861, 1862, 1863, 1864, 1866, 1867, 1870, 1877, 1879, 1882, 1886, 1888, 1889, 1890, 1892, 1896, 1897, 1899, 1901, 1907, 1909, 1915, 1919, 1920, 1925, 1926, 1928, 1930, 1931, 1932, 1933, 1935, 1936, 1940, 1941, 1944, 1947, 1948, 1949, 1950, 1953, 1955, 1958, 1959, 1961, 1963, 1964, 1971, 1973, 1974, 1978, 1979, 1981, 1983, 1984, 1988, 1989, 1990, 1991, 1992, 1994, 1995, 1996, 1997, 1998, 2000, 2001, 2006, 2013, 2015, 2017, 2020, 2025, 2029, 2034, 2035, 2036, 2038, 2039, 2040, 2041, 2042, 2044, 2046, 2047, 2048, 2049, 2052, 2054, 2058, 2059, 2062, 2063, 2064, 2070, 2071, 2072, 2073, 2074, 2077, 2081, 2085, 2087, 2088, 2099, 2100, 2102, 2106, 2107, 2109, 2110, 2114, 2118, 2119, 2121, 2123, 2124, 2126, 2129, 2130, 2131, 2132, 2133, 2134, 2138, 2139, 2140, 2141, 2142, 2146, 2147, 2149, 2152, 2156, 2161, 2162, 2163, 2164, 2167, 2169, 2171, 2173, 2176, 2177, 2178, 2180, 2181, 2182, 2184, 2186, 2188, 2189, 2190, 2194, 2195, 2196, 2197, 2198, 2199, 2204, 2205, 2208, 2212, 2215, 2219, 2221, 2222, 2223, 2226, 2230, 2233, 2237, 2240, 2241, 2242, 2243, 2245, 2246, 2247, 2248, 2250, 2251, 2256, 2257, 2260, 2261, 2262, 2265, 2267, 2274, 2275, 2278, 2279, 2280, 2283, 2285, 2287, 2289, 2292, 2297, 2298, 2301, 2306, 2307, 2311, 2312, 2313, 2320, 2321, 2322, 2324, 2325, 2326, 2330, 2333, 2335, 2337, 2338, 2339, 2340, 2343, 2344, 2350, 2351, 2359, 2360, 2361, 2362, 2364, 2365, 2366, 2370, 2372, 2378, 2379, 2388, 2395, 2400, 2404, 2406, 2408, 2410, 2412, 2420, 2422, 2423, 2424, 2425, 2426, 2429, 2437, 2440, 2442, 2444, 2445, 2446, 2448, 2450, 2451, 2453, 2454, 2457, 2459, 2461, 2463, 2464, 2465, 2466, 2469, 2470, 2472, 2475, 2477, 2479, 2480, 2481, 2484, 2488, 2491, 2494, 2495, 2496, 2497, 2500, 2501, 2508, 2509, 2511, 2512, 2513, 2515, 2516, 2517, 2520, 2522, 2523, 2529, 2530, 2532, 2535, 2537, 2538, 2539, 2540, 2544, 2547, 2548, 2549, 2550, 2553, 2554, 2555, 2556, 2558, 2559, 2560, 2562, 2566, 2570, 2571, 2574, 2575, 2576, 2577, 2586, 2587, 2589, 2590, 2592, 2594, 2602, 2604, 2605, 2606, 2607, 2609, 2610, 2619, 2622, 2623, 2624, 2625, 2629, 2631, 2632, 2633, 2635, 2636, 2637, 2638, 2641, 2644, 2647, 2649, 2651, 2652, 2655, 2657, 2660, 2663, 2665, 2667, 2675, 2676, 2678, 2680, 2681, 2684, 2685, 2686, 2687, 2688, 2691, 2694, 2695, 2697, 2702, 2703, 2707, 2708, 2711, 2713, 2714, 2716, 2717, 2718, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2733, 2735, 2738, 2740, 2742, 2744, 2746, 2749, 2754, 2757, 2760, 2762, 2764, 2766, 2767, 2768, 2769, 2771, 2774, 2776, 2777, 2778, 2783, 2786, 2787, 2788, 2789, 2791, 2793, 2798, 2801, 2802, 2803, 2810, 2811, 2814, 2816, 2817, 2820, 2821, 2824, 2829, 2831, 2836, 2837, 2841, 2843, 2845, 2846, 2849, 2850, 2851, 2857, 2858, 2860, 2863, 2865, 2867, 2868, 2870, 2873, 2874, 2877, 2878, 2879, 2880, 2881, 2882, 2884, 2887, 2888, 2892, 2896, 2897, 2899, 2902, 2904, 2905, 2912, 2916, 2920, 2921, 2922, 2925, 2927, 2930, 2932, 2933, 2937, 2939, 2940, 2942, 2944, 2946, 2949, 2950, 2951, 2953, 2954, 2955, 2956, 2958, 2960, 2964, 2967, 2968, 2971, 2972, 2973, 2975, 2976, 2978, 2980, 2981, 2982, 2983, 2985, 2994, 2996, 2998, 2999]}
+[1 rows x 10 columns]
+{'accuracy': 0.542, 'incorrect_ids': [5, 7, 9, 12, 17, 24, 26, 27, 28, 31, 32, 35, 37, 52, 55, 58, 59, 65, 66, 67, 71, 72, 76, 77, 78, 83, 84, 88, 91, 94, 95, 97, 99, 101, 104, 105, 106, 107, 108, 109, 112, 115, 117, 118, 121, 122, 123, 127, 129, 131, 135, 136, 138, 139, 144, 149, 150, 153, 155, 160, 162, 163, 164, 169, 170, 173, 174, 175, 178, 179, 180, 181, 184, 188, 189, 190, 191, 193, 194, 199, 200, 202, 203, 204, 207, 209, 214, 216, 221, 222, 223, 224, 226, 227, 228, 233, 234, 236, 237, 240, 242, 243, 245, 250, 251, 252, 253, 257, 260, 261, 262, 263, 268, 269, 271, 272, 273, 275, 278, 284, 286, 289, 293, 294, 303, 304, 305, 307, 308, 314, 315, 316, 317, 321, 322, 326, 327, 330, 332, 334, 335, 338, 339, 340, 341, 344, 347, 348, 350, 351, 352, 354, 355, 356, 360, 364, 367, 368, 371, 372, 373, 374, 376, 378, 386, 387, 389, 396, 397, 402, 404, 407, 410, 422, 423, 425, 428, 430, 431, 436, 437, 438, 440, 442, 445, 447, 450, 452, 454, 455, 456, 458, 459, 460, 464, 469, 471, 472, 476, 480, 483, 486, 488, 490, 492, 493, 495, 496, 497, 498, 500, 501, 502, 506, 507, 510, 511, 513, 514, 516, 517, 518, 519, 530, 531, 536, 538, 540, 543, 550, 552, 556, 560, 566, 568, 570, 571, 579, 581, 589, 591, 592, 593, 594, 600, 601, 612, 613, 614, 620, 621, 625, 629, 632, 636, 637, 643, 644, 646, 647, 653, 663, 665, 666, 671, 673, 674, 675, 680, 682, 683, 686, 689, 690, 691, 694, 695, 697, 702, 703, 704, 705, 706, 708, 709, 711, 715, 716, 717, 718, 720, 721, 722, 726, 729, 730, 733, 734, 736, 739, 740, 747, 752, 754, 756, 758, 759, 762, 763, 765, 766, 770, 772, 774, 777, 782, 784, 785, 786, 790, 791, 794, 795, 798, 799, 800, 801, 803, 805, 809, 812, 813, 814, 816, 817, 818, 820, 821, 822, 823, 824, 827, 828, 829, 832, 833, 834, 837, 840, 841, 842, 843, 844, 845, 847, 848, 850, 851, 856, 857, 861, 862, 866, 869, 870, 871, 873, 878, 880, 886, 887, 888, 889, 890, 891, 892, 893, 895, 896, 897, 899, 901, 904, 906, 907, 910, 913, 914, 916, 917, 920, 926, 927, 932, 933, 936, 939, 940, 945, 949, 952, 953, 954, 956, 959, 961, 962, 966, 968, 969, 970, 971, 973, 980, 982, 983, 984, 986, 987, 989, 991, 992, 993, 997, 998, 1000, 1001, 1003, 1005, 1006, 1009, 1011, 1012, 1014, 1015, 1020, 1022, 1023, 1025, 1028, 1031, 1036, 1038, 1039, 1040, 1041, 1043, 1044, 1046, 1049, 1051, 1053, 1054, 1058, 1060, 1061, 1063, 1064, 1066, 1071, 1073, 1074, 1075, 1077, 1080, 1081, 1082, 1083, 1086, 1087, 1089, 1093, 1095, 1096, 1098, 1102, 1106, 1107, 1110, 1111, 1114, 1116, 1120, 1121, 1122, 1125, 1128, 1134, 1135, 1136, 1138, 1140, 1141, 1143, 1144, 1146, 1149, 1151, 1161, 1164, 1166, 1168, 1169, 1170, 1171, 1172, 1175, 1177, 1178, 1179, 1182, 1183, 1184, 1185, 1186, 1188, 1192, 1196, 1199, 1202, 1203, 1204, 1205, 1206, 1207, 1209, 1212, 1216, 1219, 1220, 1221, 1224, 1228, 1230, 1231, 1234, 1235, 1237, 1238, 1239, 1241, 1242, 1244, 1245, 1251, 1252, 1256, 1259, 1264, 1266, 1267, 1269, 1270, 1271, 1272, 1274, 1278, 1281, 1282, 1285, 1286, 1288, 1289, 1290, 1292, 1293, 1299, 1300, 1301, 1304, 1305, 1306, 1308, 1311, 1312, 1313, 1314, 1315, 1317, 1318, 1319, 1320, 1321, 1322, 1323, 1324, 1329, 1331, 1334, 1342, 1344, 1346, 1348, 1349, 1353, 1354, 1356, 1358, 1359, 1360, 1362, 1364, 1367, 1368, 1370, 1371, 1374, 1377, 1378, 1379, 1385, 1387, 1389, 1390, 1392, 1395, 1402, 1405, 1406, 1407, 1408, 1409, 1412, 1413, 1416, 1419, 1420, 1422, 1424, 1426, 1430, 1431, 1432, 1438, 1440, 1441, 1443, 1444, 1448, 1450, 1451, 1452, 1453, 1454, 1455, 1457, 1459, 1462, 1463, 1467, 1469, 1470, 1473, 1476, 1477, 1478, 1482, 1487, 1488, 1490, 1492, 1493, 1495, 1496, 1499, 1500, 1502, 1504, 1506, 1507, 1508, 1510, 1515, 1516, 1517, 1518, 1519, 1520, 1524, 1525, 1528, 1529, 1532, 1533, 1534, 1536, 1538, 1542, 1545, 1546, 1547, 1549, 1550, 1551, 1558, 1559, 1560, 1561, 1562, 1567, 1568, 1573, 1574, 1577, 1578, 1580, 1581, 1584, 1585, 1586, 1589, 1590, 1591, 1593, 1594, 1598, 1601, 1602, 1603, 1604, 1605, 1606, 1607, 1608, 1611, 1616, 1618, 1619, 1620, 1622, 1625, 1627, 1631, 1632, 1633, 1636, 1637, 1638, 1640, 1641, 1645, 1647, 1654, 1655, 1658, 1659, 1663, 1668, 1669, 1671, 1672, 1673, 1674, 1676, 1677, 1678, 1679, 1682, 1683, 1684, 1685, 1686, 1687, 1689, 1690, 1691, 1692, 1694, 1695, 1697, 1698, 1700, 1710, 1712, 1716, 1717, 1718, 1721, 1722, 1726, 1727, 1730, 1734, 1736, 1737, 1738, 1739, 1740, 1743, 1747, 1750, 1751, 1752, 1753, 1754, 1755, 1756, 1757, 1759, 1762, 1763, 1765, 1766, 1767, 1770, 1771, 1773, 1778, 1779, 1780, 1781, 1785, 1786, 1787, 1788, 1791, 1792, 1796, 1798, 1799, 1804, 1805, 1809, 1810, 1811, 1812, 1813, 1816, 1818, 1820, 1824, 1827, 1834, 1835, 1837, 1839, 1840, 1843, 1848, 1857, 1858, 1859, 1861, 1863, 1864, 1866, 1870, 1872, 1873, 1877, 1879, 1882, 1886, 1888, 1889, 1890, 1892, 1894, 1897, 1899, 1901, 1905, 1909, 1912, 1915, 1919, 1920, 1924, 1926, 1928, 1930, 1932, 1933, 1934, 1935, 1936, 1940, 1941, 1943, 1944, 1947, 1948, 1949, 1950, 1953, 1958, 1959, 1961, 1963, 1964, 1968, 1970, 1971, 1973, 1974, 1978, 1979, 1981, 1982, 1984, 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 2000, 2001, 2002, 2010, 2013, 2017, 2020, 2025, 2026, 2029, 2035, 2038, 2040, 2041, 2042, 2044, 2046, 2047, 2048, 2052, 2053, 2054, 2058, 2059, 2062, 2064, 2068, 2070, 2072, 2073, 2074, 2077, 2078, 2079, 2081, 2082, 2083, 2085, 2087, 2088, 2091, 2099, 2100, 2101, 2102, 2104, 2106, 2107, 2109, 2110, 2112, 2114, 2118, 2119, 2121, 2122, 2124, 2126, 2129, 2132, 2133, 2135, 2139, 2141, 2142, 2143, 2145, 2146, 2147, 2149, 2150, 2152, 2156, 2161, 2162, 2164, 2167, 2169, 2171, 2173, 2176, 2177, 2179, 2180, 2182, 2184, 2185, 2186, 2187, 2190, 2192, 2193, 2194, 2196, 2197, 2198, 2199, 2205, 2206, 2208, 2209, 2212, 2215, 2219, 2221, 2222, 2226, 2227, 2230, 2233, 2238, 2241, 2242, 2243, 2245, 2246, 2247, 2248, 2250, 2251, 2253, 2257, 2260, 2261, 2262, 2263, 2265, 2266, 2267, 2274, 2275, 2278, 2279, 2280, 2282, 2283, 2285, 2287, 2289, 2292, 2293, 2297, 2298, 2301, 2306, 2311, 2313, 2319, 2320, 2321, 2323, 2324, 2325, 2333, 2334, 2335, 2337, 2339, 2344, 2349, 2350, 2351, 2359, 2360, 2364, 2365, 2366, 2367, 2369, 2370, 2372, 2376, 2378, 2379, 2380, 2382, 2386, 2388, 2390, 2393, 2395, 2398, 2399, 2400, 2401, 2404, 2405, 2406, 2408, 2410, 2412, 2420, 2423, 2424, 2425, 2426, 2437, 2440, 2442, 2444, 2445, 2446, 2448, 2450, 2453, 2454, 2459, 2461, 2463, 2464, 2465, 2466, 2469, 2470, 2472, 2475, 2476, 2477, 2481, 2484, 2488, 2491, 2494, 2496, 2497, 2499, 2500, 2501, 2503, 2508, 2511, 2512, 2513, 2516, 2517, 2518, 2520, 2522, 2523, 2528, 2529, 2530, 2532, 2534, 2535, 2538, 2539, 2540, 2544, 2546, 2547, 2548, 2549, 2550, 2554, 2556, 2558, 2559, 2560, 2562, 2566, 2570, 2571, 2574, 2575, 2576, 2577, 2587, 2589, 2590, 2592, 2594, 2604, 2605, 2606, 2607, 2609, 2610, 2612, 2616, 2619, 2623, 2624, 2625, 2629, 2631, 2633, 2635, 2636, 2637, 2641, 2643, 2644, 2647, 2649, 2653, 2654, 2655, 2657, 2658, 2660, 2661, 2663, 2665, 2667, 2669, 2670, 2672, 2675, 2676, 2677, 2678, 2681, 2684, 2685, 2686, 2687, 2688, 2691, 2694, 2695, 2697, 2702, 2703, 2704, 2707, 2708, 2711, 2713, 2714, 2715, 2716, 2717, 2718, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2733, 2738, 2740, 2742, 2744, 2745, 2746, 2751, 2754, 2756, 2757, 2760, 2762, 2764, 2766, 2767, 2768, 2769, 2771, 2774, 2776, 2777, 2778, 2781, 2786, 2787, 2788, 2791, 2793, 2795, 2798, 2801, 2803, 2805, 2810, 2811, 2816, 2817, 2820, 2821, 2822, 2824, 2829, 2831, 2836, 2837, 2838, 2841, 2843, 2845, 2849, 2851, 2854, 2857, 2858, 2860, 2861, 2862, 2863, 2865, 2867, 2870, 2873, 2874, 2877, 2878, 2879, 2880, 2881, 2882, 2884, 2885, 2887, 2892, 2896, 2897, 2899, 2902, 2904, 2905, 2910, 2912, 2915, 2916, 2919, 2921, 2930, 2933, 2938, 2939, 2940, 2942, 2944, 2946, 2949, 2950, 2953, 2954, 2955, 2956, 2957, 2958, 2960, 2962, 2964, 2965, 2968, 2971, 2972, 2975, 2976, 2978, 2979, 2980, 2981, 2982, 2983, 2985, 2986, 2994, 2998, 2999]}

logs/glm-4-9b-chat-1m_epoch_3.txt

@@ -150,8 +150,8 @@ step 4: 不是
 step 5: 不是
 (3) GPU = NVIDIA L40. Max memory = 44.309 GB.
 18.473 GB of memory reserved.
-      text  ... THUDM/glm-4-9b-chat-1m_checkpoint-525
+      text  ... THUDM/glm-4-9b-chat-1m_checkpoint-700
 0  甄加索是自杀吗  ...                                    不是
 
-[1 rows x 9 columns]
-{'accuracy': 0.5986666666666667, 'incorrect_ids': [7, 9, 12, 13, 14, 17, 24, 25, 26, 27, 28, 31, 33, 35, 36, 38, 52, 55, 58, 59, 65, 67, 71, 72, 76, 77, 78, 81, 83, 84, 88, 93, 96, 99, 101, 104, 105, 108, 109, 112, 113, 115, 117, 118, 121, 123, 124, 127, 131, 135, 136, 137, 138, 139, 143, 144, 150, 153, 154, 155, 160, 161, 164, 171, 172, 174, 175, 178, 179, 180, 181, 187, 188, 190, 191, 192, 193, 197, 199, 200, 202, 204, 207, 214, 215, 218, 220, 223, 224, 225, 226, 227, 228, 230, 233, 234, 236, 240, 243, 245, 248, 250, 252, 253, 256, 257, 260, 262, 263, 265, 269, 271, 272, 273, 275, 276, 283, 284, 286, 289, 301, 303, 304, 305, 307, 308, 313, 314, 315, 316, 317, 320, 321, 330, 334, 335, 338, 339, 340, 341, 344, 348, 349, 350, 353, 354, 355, 356, 358, 360, 367, 368, 371, 373, 374, 376, 378, 386, 389, 395, 396, 397, 402, 403, 404, 410, 414, 422, 423, 428, 429, 430, 431, 436, 438, 445, 447, 450, 452, 454, 456, 457, 459, 460, 461, 464, 471, 472, 475, 476, 480, 482, 483, 488, 490, 492, 494, 496, 498, 500, 501, 502, 503, 506, 507, 508, 511, 513, 514, 518, 519, 520, 536, 538, 553, 560, 566, 570, 571, 576, 579, 581, 585, 589, 591, 592, 593, 597, 601, 612, 613, 614, 621, 625, 627, 628, 632, 636, 637, 643, 644, 646, 647, 649, 665, 666, 671, 673, 674, 675, 676, 682, 683, 686, 689, 690, 691, 694, 695, 697, 701, 702, 703, 706, 709, 717, 718, 721, 726, 729, 730, 732, 734, 739, 740, 742, 747, 752, 754, 757, 758, 766, 770, 771, 772, 773, 774, 778, 779, 782, 784, 785, 788, 791, 795, 798, 799, 800, 801, 805, 809, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 826, 827, 828, 833, 834, 841, 843, 845, 847, 848, 850, 851, 853, 856, 857, 860, 861, 866, 869, 870, 871, 873, 876, 878, 884, 886, 888, 889, 890, 891, 893, 895, 896, 897, 899, 901, 904, 907, 910, 913, 914, 920, 926, 927, 932, 933, 935, 936, 937, 939, 940, 942, 945, 948, 949, 952, 953, 954, 956, 959, 961, 962, 966, 969, 970, 973, 974, 980, 981, 983, 984, 987, 988, 991, 994, 998, 1001, 1003, 1006, 1007, 1009, 1011, 1012, 1014, 1015, 1022, 1023, 1025, 1026, 1028, 1036, 1038, 1039, 1043, 1044, 1045, 1046, 1048, 1051, 1054, 1056, 1058, 1063, 1069, 1071, 1073, 1074, 1077, 1080, 1081, 1084, 1087, 1089, 1091, 1093, 1096, 1098, 1102, 1106, 1107, 1108, 1110, 1111, 1114, 1116, 1120, 1121, 1124, 1125, 1126, 1128, 1129, 1134, 1135, 1136, 1140, 1143, 1156, 1158, 1160, 1161, 1162, 1163, 1164, 1166, 1167, 1168, 1171, 1172, 1177, 1178, 1180, 1181, 1185, 1192, 1196, 1202, 1203, 1206, 1209, 1212, 1216, 1219, 1220, 1224, 1227, 1228, 1229, 1230, 1231, 1235, 1237, 1238, 1239, 1241, 1242, 1245, 1251, 1252, 1256, 1259, 1264, 1266, 1267, 1270, 1271, 1272, 1273, 1274, 1281, 1282, 1286, 1289, 1290, 1292, 1293, 1299, 1300, 1301, 1304, 1305, 1308, 1312, 1313, 1316, 1317, 1318, 1322, 1324, 1329, 1331, 1342, 1344, 1346, 1348, 1349, 1353, 1359, 1360, 1363, 1365, 1370, 1371, 1372, 1373, 1378, 1379, 1384, 1385, 1386, 1387, 1389, 1392, 1393, 1395, 1400, 1402, 1406, 1407, 1409, 1413, 1418, 1419, 1420, 1423, 1424, 1425, 1426, 1430, 1431, 1432, 1438, 1440, 1441, 1443, 1444, 1450, 1452, 1453, 1454, 1455, 1457, 1459, 1462, 1469, 1473, 1477, 1478, 1481, 1482, 1484, 1487, 1488, 1489, 1490, 1491, 1495, 1496, 1502, 1503, 1504, 1507, 1508, 1510, 1512, 1516, 1517, 1518, 1525, 1528, 1533, 1535, 1536, 1542, 1544, 1545, 1546, 1547, 1549, 1551, 1555, 1558, 1560, 1561, 1562, 1568, 1574, 1580, 1581, 1585, 1587, 1590, 1591, 1593, 1594, 1600, 1602, 1603, 1604, 1605, 1606, 1607, 1608, 1611, 1612, 1613, 1619, 1622, 1627, 1631, 1632, 1633, 1636, 1637, 1639, 1641, 1647, 1648, 1650, 1654, 1655, 1657, 1659, 1660, 1662, 1665, 1671, 1672, 1674, 1677, 1678, 1679, 1680, 1682, 1683, 1684, 1686, 1691, 1694, 1695, 1697, 1698, 1704, 1710, 1711, 1713, 1716, 1718, 1720, 1721, 1722, 1726, 1729, 1733, 1734, 1735, 1736, 1737, 1738, 1740, 1743, 1745, 1751, 1753, 1754, 1755, 1756, 1758, 1759, 1765, 1766, 1767, 1769, 1770, 1771, 1778, 1779, 1780, 1783, 1785, 1786, 1788, 1790, 1791, 1798, 1799, 1800, 1805, 1809, 1810, 1812, 1816, 1818, 1824, 1827, 1834, 1835, 1839, 1848, 1853, 1854, 1855, 1856, 1857, 1859, 1860, 1862, 1866, 1867, 1869, 1873, 1877, 1879, 1880, 1882, 1888, 1890, 1891, 1892, 1897, 1900, 1913, 1914, 1915, 1919, 1920, 1923, 1928, 1931, 1933, 1934, 1935, 1936, 1938, 1940, 1941, 1946, 1948, 1949, 1951, 1953, 1955, 1958, 1959, 1961, 1962, 1963, 1964, 1970, 1971, 1973, 1974, 1978, 1981, 1984, 1988, 1989, 1990, 1992, 1994, 1995, 1998, 2000, 2001, 2008, 2013, 2014, 2015, 2017, 2019, 2020, 2022, 2028, 2029, 2034, 2035, 2036, 2038, 2041, 2042, 2044, 2046, 2048, 2049, 2052, 2053, 2054, 2056, 2058, 2062, 2063, 2064, 2070, 2071, 2077, 2081, 2085, 2087, 2088, 2092, 2099, 2100, 2106, 2109, 2110, 2112, 2114, 2118, 2119, 2121, 2122, 2124, 2126, 2131, 2132, 2133, 2139, 2141, 2142, 2145, 2146, 2147, 2152, 2156, 2159, 2160, 2161, 2162, 2167, 2169, 2172, 2173, 2177, 2180, 2181, 2184, 2185, 2186, 2187, 2188, 2190, 2192, 2193, 2194, 2196, 2197, 2198, 2199, 2200, 2203, 2212, 2213, 2214, 2215, 2219, 2222, 2223, 2226, 2230, 2231, 2233, 2240, 2241, 2242, 2243, 2245, 2246, 2247, 2248, 2250, 2251, 2261, 2262, 2263, 2265, 2267, 2274, 2278, 2279, 2280, 2287, 2289, 2290, 2294, 2295, 2297, 2298, 2301, 2306, 2311, 2312, 2313, 2314, 2320, 2324, 2331, 2333, 2335, 2337, 2338, 2339, 2343, 2344, 2348, 2351, 2352, 2353, 2360, 2364, 2366, 2367, 2376, 2378, 2379, 2388, 2389, 2390, 2395, 2396, 2398, 2400, 2404, 2405, 2406, 2408, 2410, 2412, 2420, 2421, 2423, 2425, 2426, 2437, 2440, 2442, 2444, 2445, 2448, 2450, 2453, 2454, 2463, 2464, 2465, 2466, 2470, 2472, 2473, 2475, 2476, 2479, 2484, 2488, 2491, 2493, 2495, 2496, 2497, 2499, 2500, 2501, 2502, 2503, 2507, 2509, 2510, 2511, 2513, 2515, 2517, 2518, 2520, 2522, 2525, 2526, 2529, 2530, 2535, 2537, 2538, 2539, 2540, 2544, 2547, 2548, 2549, 2551, 2555, 2556, 2558, 2559, 2560, 2562, 2563, 2570, 2571, 2574, 2575, 2576, 2577, 2586, 2589, 2590, 2592, 2593, 2594, 2602, 2605, 2606, 2607, 2609, 2620, 2622, 2624, 2625, 2628, 2629, 2630, 2631, 2633, 2638, 2641, 2642, 2644, 2645, 2647, 2653, 2657, 2660, 2663, 2664, 2667, 2672, 2675, 2678, 2684, 2685, 2687, 2688, 2691, 2694, 2697, 2702, 2703, 2707, 2713, 2714, 2716, 2717, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2733, 2738, 2742, 2744, 2745, 2746, 2747, 2752, 2754, 2756, 2757, 2759, 2760, 2761, 2762, 2763, 2764, 2766, 2767, 2768, 2769, 2771, 2774, 2776, 2777, 2778, 2781, 2788, 2791, 2792, 2793, 2795, 2798, 2800, 2801, 2805, 2809, 2810, 2811, 2813, 2816, 2820, 2821, 2824, 2829, 2836, 2837, 2840, 2841, 2843, 2845, 2850, 2851, 2854, 2857, 2858, 2860, 2861, 2863, 2870, 2871, 2873, 2875, 2877, 2878, 2879, 2880, 2881, 2882, 2884, 2892, 2896, 2897, 2899, 2902, 2904, 2905, 2910, 2912, 2915, 2916, 2921, 2930, 2933, 2935, 2937, 2938, 2939, 2942, 2944, 2946, 2949, 2951, 2954, 2955, 2956, 2960, 2963, 2964, 2968, 2969, 2971, 2972, 2976, 2978, 2979, 2980, 2981, 2982, 2985, 2991, 2994, 2998, 2999]}
+[1 rows x 10 columns]
+{'accuracy': 0.604, 'incorrect_ids': [7, 9, 12, 13, 17, 24, 25, 26, 27, 28, 31, 33, 34, 36, 37, 38, 52, 53, 55, 59, 65, 67, 71, 72, 76, 77, 78, 81, 83, 84, 88, 91, 93, 96, 99, 101, 103, 104, 105, 108, 109, 112, 113, 115, 117, 118, 121, 127, 131, 135, 136, 137, 138, 139, 143, 144, 150, 153, 155, 160, 161, 164, 169, 171, 174, 175, 178, 179, 180, 181, 184, 187, 188, 190, 191, 192, 193, 197, 199, 200, 202, 207, 214, 215, 221, 223, 224, 225, 227, 228, 233, 236, 237, 240, 243, 245, 248, 250, 251, 252, 253, 256, 257, 260, 262, 263, 265, 268, 269, 271, 272, 273, 275, 276, 284, 286, 289, 293, 301, 303, 304, 305, 307, 308, 313, 314, 315, 316, 317, 320, 321, 323, 327, 328, 330, 334, 335, 337, 338, 339, 340, 341, 344, 348, 349, 350, 353, 354, 355, 356, 360, 367, 368, 371, 372, 373, 374, 378, 386, 387, 389, 395, 396, 397, 402, 404, 410, 414, 422, 423, 428, 430, 431, 436, 438, 445, 447, 450, 452, 453, 454, 455, 456, 459, 460, 461, 465, 471, 472, 473, 475, 476, 480, 483, 488, 490, 492, 493, 494, 496, 498, 500, 501, 502, 503, 506, 507, 508, 511, 513, 514, 517, 518, 519, 520, 530, 531, 536, 540, 553, 560, 566, 570, 571, 576, 579, 581, 589, 591, 592, 593, 596, 597, 598, 601, 612, 613, 614, 621, 626, 627, 628, 632, 636, 642, 643, 644, 646, 647, 649, 653, 665, 666, 671, 673, 674, 676, 682, 683, 686, 689, 690, 691, 701, 702, 703, 706, 709, 717, 720, 721, 726, 729, 730, 732, 733, 734, 740, 741, 743, 747, 754, 758, 766, 772, 773, 774, 778, 782, 784, 788, 791, 795, 798, 799, 800, 801, 805, 809, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 826, 828, 833, 834, 840, 841, 842, 843, 845, 847, 848, 850, 853, 856, 860, 861, 866, 869, 870, 871, 873, 876, 886, 888, 889, 890, 891, 893, 894, 896, 897, 899, 901, 904, 907, 913, 914, 920, 926, 927, 932, 933, 936, 937, 939, 940, 945, 948, 949, 952, 953, 954, 956, 959, 961, 966, 969, 970, 973, 980, 981, 982, 983, 984, 988, 994, 998, 1001, 1003, 1007, 1009, 1011, 1012, 1014, 1015, 1022, 1023, 1026, 1028, 1036, 1038, 1039, 1043, 1044, 1045, 1046, 1048, 1051, 1053, 1054, 1056, 1058, 1061, 1063, 1065, 1069, 1071, 1073, 1074, 1075, 1076, 1077, 1080, 1081, 1087, 1089, 1093, 1096, 1098, 1099, 1100, 1102, 1106, 1107, 1108, 1110, 1114, 1120, 1121, 1124, 1125, 1126, 1128, 1129, 1133, 1134, 1136, 1140, 1141, 1143, 1144, 1151, 1155, 1156, 1158, 1159, 1160, 1161, 1162, 1163, 1166, 1167, 1168, 1170, 1171, 1172, 1176, 1177, 1178, 1180, 1181, 1185, 1192, 1196, 1203, 1206, 1209, 1212, 1213, 1216, 1219, 1220, 1224, 1227, 1228, 1229, 1230, 1231, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1245, 1251, 1252, 1256, 1259, 1264, 1266, 1267, 1270, 1271, 1272, 1273, 1274, 1281, 1282, 1285, 1286, 1289, 1290, 1292, 1293, 1295, 1296, 1299, 1300, 1301, 1305, 1308, 1312, 1313, 1315, 1316, 1317, 1318, 1320, 1321, 1322, 1323, 1324, 1329, 1331, 1339, 1342, 1344, 1346, 1347, 1348, 1349, 1353, 1354, 1358, 1359, 1360, 1362, 1363, 1365, 1368, 1370, 1371, 1372, 1374, 1378, 1379, 1384, 1385, 1386, 1387, 1389, 1391, 1393, 1400, 1402, 1405, 1406, 1407, 1409, 1412, 1416, 1419, 1420, 1423, 1424, 1425, 1426, 1427, 1432, 1438, 1440, 1441, 1443, 1444, 1445, 1446, 1450, 1452, 1453, 1454, 1457, 1459, 1462, 1469, 1473, 1477, 1478, 1481, 1482, 1490, 1491, 1495, 1496, 1498, 1502, 1503, 1504, 1506, 1507, 1508, 1510, 1512, 1516, 1517, 1518, 1525, 1527, 1528, 1533, 1535, 1536, 1544, 1545, 1546, 1547, 1549, 1551, 1555, 1558, 1560, 1561, 1562, 1568, 1574, 1580, 1581, 1585, 1587, 1589, 1590, 1593, 1594, 1600, 1601, 1603, 1604, 1605, 1606, 1607, 1608, 1611, 1612, 1613, 1619, 1622, 1627, 1631, 1632, 1633, 1637, 1639, 1641, 1647, 1648, 1650, 1654, 1655, 1657, 1660, 1662, 1663, 1668, 1671, 1672, 1674, 1676, 1677, 1678, 1679, 1680, 1682, 1683, 1684, 1686, 1691, 1694, 1695, 1697, 1698, 1700, 1710, 1711, 1713, 1716, 1718, 1720, 1721, 1722, 1726, 1729, 1733, 1734, 1735, 1736, 1737, 1738, 1739, 1740, 1743, 1745, 1751, 1753, 1754, 1755, 1756, 1758, 1759, 1765, 1766, 1767, 1768, 1769, 1770, 1771, 1778, 1779, 1780, 1783, 1784, 1786, 1788, 1791, 1796, 1798, 1799, 1805, 1809, 1810, 1812, 1813, 1816, 1818, 1824, 1825, 1827, 1834, 1835, 1840, 1841, 1846, 1852, 1853, 1855, 1856, 1857, 1859, 1862, 1866, 1867, 1877, 1879, 1880, 1882, 1886, 1888, 1890, 1891, 1892, 1897, 1900, 1913, 1914, 1916, 1919, 1920, 1923, 1926, 1928, 1930, 1931, 1932, 1933, 1934, 1935, 1938, 1940, 1941, 1944, 1946, 1948, 1949, 1951, 1953, 1958, 1959, 1961, 1962, 1963, 1964, 1971, 1973, 1974, 1978, 1981, 1984, 1988, 1989, 1990, 1992, 1994, 1995, 1996, 1998, 2000, 2013, 2014, 2015, 2017, 2019, 2020, 2022, 2025, 2034, 2035, 2036, 2038, 2041, 2042, 2048, 2049, 2052, 2053, 2054, 2058, 2062, 2064, 2070, 2076, 2077, 2078, 2081, 2085, 2087, 2088, 2091, 2099, 2100, 2104, 2106, 2109, 2110, 2114, 2119, 2121, 2124, 2126, 2131, 2132, 2133, 2139, 2141, 2142, 2145, 2146, 2147, 2152, 2159, 2160, 2161, 2162, 2164, 2167, 2169, 2172, 2173, 2177, 2180, 2181, 2184, 2185, 2186, 2187, 2190, 2192, 2193, 2194, 2195, 2196, 2198, 2199, 2200, 2203, 2204, 2212, 2214, 2215, 2219, 2222, 2223, 2226, 2230, 2231, 2233, 2238, 2240, 2241, 2243, 2245, 2247, 2248, 2256, 2260, 2261, 2262, 2267, 2274, 2278, 2280, 2285, 2287, 2288, 2289, 2290, 2297, 2298, 2301, 2304, 2306, 2311, 2312, 2314, 2319, 2322, 2323, 2324, 2333, 2335, 2337, 2339, 2343, 2344, 2352, 2353, 2360, 2361, 2364, 2366, 2367, 2376, 2378, 2379, 2388, 2389, 2390, 2395, 2398, 2400, 2404, 2405, 2406, 2408, 2410, 2412, 2416, 2420, 2421, 2423, 2424, 2425, 2426, 2429, 2437, 2440, 2442, 2444, 2445, 2450, 2453, 2454, 2459, 2463, 2464, 2465, 2466, 2470, 2472, 2473, 2475, 2488, 2491, 2493, 2496, 2497, 2499, 2500, 2502, 2503, 2509, 2510, 2511, 2512, 2513, 2515, 2517, 2518, 2520, 2522, 2526, 2529, 2530, 2532, 2535, 2538, 2539, 2540, 2544, 2547, 2548, 2549, 2551, 2556, 2559, 2562, 2574, 2575, 2576, 2586, 2589, 2590, 2592, 2594, 2596, 2602, 2605, 2607, 2609, 2610, 2619, 2620, 2622, 2623, 2624, 2625, 2626, 2628, 2629, 2630, 2631, 2633, 2637, 2638, 2640, 2644, 2645, 2647, 2653, 2657, 2662, 2663, 2664, 2667, 2672, 2675, 2677, 2678, 2686, 2687, 2688, 2694, 2702, 2703, 2707, 2713, 2714, 2716, 2717, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2738, 2742, 2744, 2746, 2747, 2752, 2754, 2756, 2757, 2760, 2761, 2762, 2764, 2766, 2767, 2768, 2769, 2771, 2774, 2776, 2777, 2778, 2781, 2788, 2791, 2792, 2793, 2795, 2798, 2800, 2801, 2805, 2809, 2811, 2813, 2816, 2817, 2820, 2821, 2824, 2829, 2831, 2833, 2836, 2837, 2840, 2841, 2843, 2844, 2845, 2850, 2851, 2854, 2857, 2858, 2860, 2863, 2865, 2870, 2873, 2875, 2877, 2879, 2880, 2881, 2882, 2884, 2892, 2896, 2897, 2899, 2902, 2904, 2905, 2910, 2912, 2915, 2916, 2921, 2922, 2925, 2930, 2933, 2937, 2938, 2939, 2944, 2946, 2949, 2951, 2953, 2954, 2955, 2956, 2960, 2963, 2964, 2968, 2971, 2972, 2975, 2976, 2978, 2979, 2980, 2981, 2982, 2985, 2991, 2994, 2998, 2999]}

logs/glm-4-9b-chat-1m_epoch_4.txt

@@ -154,4 +154,4 @@ step 5: 不是
 0  甄加索是自杀吗  ...                                    不是
 
 [1 rows x 10 columns]
-{'accuracy': 0.5843333333333334, 'incorrect_ids': [7, 9, 12, 13, 17, 24, 25, 26, 27, 28, 31, 32, 33, 34, 35, 36, 38, 52, 55, 58, 59, 61, 65, 67, 69, 71, 72, 76, 77, 78, 81, 83, 84, 88, 91, 93, 96, 99, 101, 103, 104, 105, 108, 109, 112, 113, 115, 117, 121, 123, 124, 127, 131, 135, 136, 139, 143, 144, 150, 153, 154, 155, 160, 161, 164, 169, 171, 173, 174, 175, 178, 179, 181, 184, 185, 187, 188, 190, 192, 193, 199, 200, 202, 207, 209, 214, 215, 220, 222, 224, 225, 227, 228, 230, 233, 234, 236, 237, 240, 243, 248, 250, 252, 253, 257, 260, 263, 268, 269, 271, 272, 273, 275, 276, 283, 284, 286, 289, 293, 294, 301, 303, 304, 305, 307, 308, 314, 315, 316, 317, 321, 323, 330, 333, 334, 335, 338, 339, 340, 344, 348, 349, 350, 353, 354, 355, 356, 358, 360, 362, 364, 367, 368, 369, 370, 371, 372, 373, 374, 377, 378, 386, 389, 396, 397, 402, 404, 414, 421, 422, 423, 425, 428, 430, 431, 436, 438, 445, 447, 450, 452, 454, 455, 456, 457, 459, 460, 461, 465, 471, 472, 473, 475, 476, 478, 480, 488, 490, 492, 493, 494, 498, 499, 500, 501, 502, 506, 507, 508, 510, 511, 513, 514, 518, 520, 530, 531, 536, 538, 553, 560, 566, 570, 571, 576, 579, 581, 589, 591, 592, 593, 594, 597, 598, 600, 601, 612, 613, 614, 621, 627, 628, 629, 632, 636, 642, 643, 644, 646, 647, 649, 663, 665, 666, 671, 673, 674, 676, 682, 683, 686, 689, 690, 691, 694, 699, 701, 702, 703, 705, 706, 716, 717, 718, 721, 722, 725, 726, 729, 730, 732, 734, 739, 740, 747, 752, 754, 758, 763, 766, 770, 771, 772, 773, 774, 778, 782, 784, 785, 791, 795, 798, 799, 800, 805, 809, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 826, 827, 828, 832, 833, 834, 841, 842, 843, 845, 847, 848, 850, 852, 853, 856, 861, 866, 869, 870, 871, 873, 884, 886, 888, 889, 890, 891, 893, 894, 895, 896, 897, 899, 901, 904, 907, 908, 913, 914, 917, 920, 926, 927, 929, 932, 933, 935, 936, 937, 939, 940, 945, 949, 952, 953, 954, 956, 959, 961, 962, 966, 968, 969, 970, 971, 973, 980, 981, 982, 984, 986, 988, 991, 993, 994, 997, 998, 1001, 1003, 1009, 1011, 1012, 1014, 1015, 1022, 1026, 1028, 1036, 1038, 1039, 1040, 1043, 1044, 1046, 1048, 1051, 1053, 1054, 1055, 1056, 1058, 1063, 1069, 1071, 1073, 1074, 1077, 1080, 1081, 1084, 1086, 1087, 1089, 1093, 1096, 1098, 1099, 1102, 1106, 1107, 1108, 1111, 1114, 1116, 1120, 1121, 1125, 1126, 1128, 1129, 1133, 1134, 1135, 1136, 1140, 1141, 1143, 1151, 1158, 1159, 1161, 1162, 1163, 1164, 1167, 1171, 1172, 1177, 1178, 1180, 1181, 1185, 1188, 1192, 1196, 1202, 1203, 1206, 1209, 1212, 1217, 1219, 1220, 1224, 1227, 1228, 1230, 1231, 1232, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1245, 1251, 1252, 1256, 1259, 1264, 1266, 1267, 1270, 1271, 1272, 1273, 1274, 1281, 1282, 1285, 1286, 1289, 1292, 1293, 1295, 1296, 1299, 1300, 1304, 1305, 1308, 1312, 1313, 1314, 1315, 1317, 1318, 1321, 1322, 1323, 1324, 1329, 1331, 1342, 1346, 1347, 1348, 1349, 1353, 1358, 1359, 1360, 1362, 1363, 1365, 1370, 1371, 1372, 1373, 1376, 1379, 1380, 1384, 1385, 1386, 1387, 1389, 1390, 1392, 1393, 1394, 1395, 1400, 1402, 1406, 1407, 1409, 1412, 1413, 1419, 1420, 1422, 1424, 1426, 1427, 1431, 1432, 1438, 1440, 1443, 1444, 1445, 1448, 1450, 1452, 1453, 1454, 1455, 1457, 1460, 1462, 1469, 1473, 1474, 1476, 1477, 1478, 1481, 1487, 1488, 1489, 1490, 1492, 1496, 1500, 1502, 1503, 1504, 1507, 1508, 1510, 1512, 1516, 1517, 1518, 1523, 1524, 1525, 1528, 1533, 1535, 1536, 1545, 1546, 1547, 1549, 1551, 1552, 1558, 1560, 1561, 1562, 1568, 1573, 1574, 1581, 1585, 1590, 1591, 1593, 1594, 1599, 1600, 1602, 1603, 1604, 1605, 1607, 1608, 1611, 1612, 1613, 1619, 1622, 1624, 1627, 1631, 1632, 1633, 1636, 1637, 1638, 1639, 1641, 1644, 1647, 1648, 1650, 1654, 1655, 1656, 1657, 1658, 1659, 1660, 1667, 1668, 1672, 1674, 1677, 1678, 1679, 1680, 1682, 1684, 1686, 1691, 1693, 1694, 1695, 1697, 1701, 1704, 1710, 1711, 1713, 1716, 1717, 1718, 1720, 1721, 1722, 1726, 1729, 1733, 1734, 1735, 1736, 1737, 1738, 1739, 1740, 1743, 1745, 1747, 1751, 1753, 1754, 1755, 1756, 1758, 1759, 1765, 1767, 1769, 1770, 1771, 1772, 1778, 1779, 1780, 1783, 1786, 1788, 1791, 1798, 1799, 1800, 1805, 1809, 1810, 1812, 1816, 1818, 1823, 1824, 1825, 1827, 1834, 1835, 1840, 1841, 1846, 1848, 1849, 1852, 1853, 1854, 1855, 1856, 1857, 1859, 1862, 1863, 1864, 1866, 1867, 1869, 1873, 1877, 1879, 1880, 1882, 1888, 1889, 1890, 1892, 1897, 1900, 1909, 1913, 1919, 1920, 1926, 1928, 1930, 1931, 1933, 1934, 1935, 1936, 1938, 1940, 1941, 1944, 1948, 1949, 1951, 1953, 1955, 1958, 1959, 1961, 1962, 1963, 1964, 1965, 1970, 1971, 1973, 1978, 1979, 1981, 1984, 1988, 1989, 1990, 1992, 1994, 1995, 1998, 2000, 2005, 2008, 2013, 2014, 2015, 2017, 2019, 2020, 2025, 2029, 2030, 2035, 2036, 2038, 2039, 2041, 2042, 2046, 2048, 2049, 2052, 2054, 2058, 2063, 2064, 2070, 2071, 2072, 2077, 2078, 2079, 2081, 2082, 2085, 2087, 2088, 2090, 2091, 2092, 2095, 2099, 2100, 2106, 2109, 2110, 2112, 2114, 2118, 2119, 2121, 2122, 2124, 2126, 2131, 2132, 2133, 2135, 2138, 2139, 2141, 2142, 2145, 2146, 2147, 2151, 2152, 2159, 2160, 2161, 2162, 2164, 2167, 2169, 2172, 2173, 2177, 2179, 2180, 2181, 2184, 2185, 2186, 2187, 2190, 2192, 2193, 2194, 2195, 2196, 2198, 2199, 2200, 2203, 2212, 2215, 2219, 2222, 2223, 2226, 2230, 2231, 2233, 2240, 2241, 2242, 2243, 2245, 2246, 2247, 2248, 2249, 2250, 2253, 2256, 2260, 2261, 2262, 2263, 2267, 2274, 2278, 2279, 2280, 2289, 2290, 2293, 2297, 2298, 2301, 2304, 2306, 2311, 2312, 2313, 2318, 2319, 2320, 2322, 2324, 2333, 2335, 2337, 2338, 2339, 2342, 2343, 2344, 2353, 2359, 2360, 2361, 2364, 2366, 2367, 2369, 2370, 2376, 2378, 2379, 2382, 2388, 2389, 2390, 2395, 2398, 2400, 2404, 2405, 2406, 2408, 2409, 2410, 2412, 2416, 2420, 2421, 2423, 2425, 2426, 2437, 2440, 2441, 2442, 2444, 2445, 2448, 2450, 2453, 2454, 2461, 2463, 2464, 2465, 2470, 2472, 2473, 2474, 2475, 2477, 2479, 2481, 2484, 2488, 2491, 2493, 2496, 2497, 2499, 2500, 2501, 2507, 2510, 2511, 2512, 2513, 2515, 2517, 2518, 2522, 2525, 2529, 2530, 2532, 2535, 2537, 2538, 2539, 2540, 2544, 2546, 2547, 2548, 2549, 2551, 2553, 2556, 2557, 2559, 2562, 2570, 2575, 2576, 2577, 2578, 2583, 2586, 2589, 2590, 2592, 2593, 2594, 2602, 2605, 2606, 2607, 2609, 2610, 2616, 2620, 2622, 2623, 2624, 2625, 2626, 2628, 2629, 2630, 2631, 2633, 2636, 2638, 2641, 2644, 2645, 2647, 2653, 2657, 2663, 2664, 2665, 2667, 2672, 2676, 2677, 2678, 2687, 2688, 2694, 2701, 2702, 2703, 2707, 2708, 2713, 2714, 2716, 2717, 2718, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2733, 2738, 2742, 2744, 2745, 2746, 2752, 2754, 2756, 2757, 2758, 2759, 2760, 2761, 2762, 2763, 2764, 2765, 2766, 2768, 2769, 2771, 2772, 2774, 2776, 2777, 2781, 2788, 2791, 2792, 2793, 2795, 2798, 2800, 2801, 2803, 2810, 2813, 2816, 2820, 2821, 2824, 2829, 2836, 2837, 2841, 2843, 2845, 2850, 2851, 2854, 2857, 2858, 2860, 2863, 2864, 2867, 2870, 2871, 2873, 2875, 2877, 2878, 2879, 2880, 2881, 2882, 2884, 2892, 2896, 2897, 2898, 2899, 2902, 2905, 2910, 2912, 2915, 2916, 2920, 2921, 2922, 2925, 2929, 2930, 2933, 2934, 2935, 2937, 2938, 2939, 2940, 2942, 2944, 2946, 2949, 2954, 2955, 2956, 2957, 2960, 2963, 2964, 2968, 2971, 2972, 2975, 2976, 2978, 2979, 2980, 2981, 2982, 2985, 2991, 2994, 2995, 2998]}
+{'accuracy': 0.5813333333333334, 'incorrect_ids': [7, 9, 12, 13, 17, 24, 25, 26, 27, 28, 31, 32, 33, 34, 35, 36, 38, 52, 55, 58, 59, 65, 66, 67, 71, 72, 76, 77, 78, 81, 83, 84, 91, 93, 96, 99, 101, 103, 104, 105, 106, 108, 109, 112, 113, 114, 115, 117, 121, 123, 124, 127, 131, 135, 136, 139, 143, 144, 150, 152, 153, 154, 155, 160, 161, 164, 169, 171, 172, 173, 174, 175, 178, 184, 185, 187, 188, 190, 192, 193, 199, 200, 202, 207, 214, 215, 222, 224, 225, 226, 227, 228, 230, 233, 234, 236, 237, 240, 243, 245, 248, 250, 251, 252, 253, 257, 260, 263, 268, 271, 272, 273, 275, 276, 283, 284, 286, 289, 293, 294, 301, 303, 304, 305, 307, 308, 311, 314, 315, 316, 317, 320, 321, 323, 330, 333, 334, 335, 337, 338, 339, 340, 344, 348, 349, 350, 353, 354, 355, 356, 357, 358, 360, 362, 364, 367, 368, 371, 372, 373, 374, 378, 386, 389, 395, 396, 397, 402, 404, 414, 421, 422, 423, 425, 428, 430, 431, 436, 438, 445, 447, 450, 452, 453, 454, 455, 456, 457, 459, 460, 461, 464, 465, 471, 472, 474, 475, 476, 480, 482, 488, 490, 492, 493, 494, 498, 499, 500, 501, 502, 506, 507, 508, 511, 513, 514, 518, 519, 520, 530, 531, 536, 538, 553, 560, 566, 570, 571, 576, 579, 581, 589, 591, 592, 593, 596, 597, 599, 601, 612, 613, 614, 621, 625, 627, 628, 632, 636, 642, 643, 644, 646, 647, 649, 653, 663, 665, 666, 671, 673, 674, 675, 676, 682, 683, 686, 689, 690, 691, 694, 697, 699, 701, 702, 703, 705, 706, 716, 717, 718, 719, 720, 721, 722, 729, 730, 732, 734, 739, 740, 747, 752, 754, 758, 763, 766, 770, 771, 772, 773, 774, 778, 782, 784, 785, 788, 791, 795, 798, 799, 800, 805, 809, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 826, 827, 828, 832, 833, 834, 840, 841, 842, 843, 845, 847, 848, 850, 852, 853, 856, 860, 861, 866, 869, 870, 871, 873, 878, 884, 886, 888, 889, 890, 891, 893, 894, 895, 896, 897, 899, 901, 904, 906, 907, 908, 913, 914, 920, 926, 927, 933, 935, 936, 937, 939, 940, 942, 945, 949, 952, 953, 954, 956, 959, 961, 962, 966, 968, 969, 970, 973, 974, 980, 982, 984, 988, 991, 994, 997, 998, 1001, 1003, 1007, 1011, 1012, 1014, 1015, 1022, 1025, 1026, 1028, 1036, 1038, 1039, 1040, 1044, 1045, 1046, 1048, 1051, 1053, 1054, 1055, 1056, 1058, 1063, 1069, 1071, 1073, 1074, 1075, 1077, 1080, 1081, 1084, 1086, 1087, 1089, 1093, 1095, 1096, 1098, 1099, 1102, 1106, 1107, 1108, 1110, 1111, 1114, 1116, 1120, 1121, 1123, 1125, 1126, 1128, 1129, 1133, 1134, 1136, 1138, 1140, 1141, 1143, 1144, 1151, 1156, 1158, 1160, 1161, 1162, 1163, 1164, 1166, 1167, 1171, 1172, 1174, 1177, 1178, 1180, 1181, 1182, 1185, 1192, 1196, 1202, 1203, 1204, 1206, 1209, 1212, 1216, 1217, 1219, 1220, 1224, 1227, 1228, 1230, 1231, 1232, 1235, 1237, 1238, 1239, 1240, 1241, 1242, 1245, 1249, 1251, 1252, 1256, 1259, 1264, 1266, 1267, 1269, 1270, 1271, 1272, 1274, 1281, 1282, 1285, 1286, 1289, 1292, 1293, 1295, 1296, 1299, 1300, 1304, 1305, 1308, 1311, 1312, 1313, 1314, 1315, 1317, 1318, 1322, 1324, 1329, 1331, 1332, 1342, 1346, 1347, 1348, 1349, 1353, 1354, 1358, 1359, 1360, 1362, 1363, 1365, 1370, 1371, 1372, 1373, 1379, 1384, 1385, 1386, 1387, 1389, 1390, 1392, 1393, 1394, 1395, 1400, 1402, 1405, 1406, 1407, 1409, 1412, 1419, 1420, 1424, 1426, 1427, 1431, 1432, 1438, 1440, 1443, 1444, 1445, 1448, 1450, 1452, 1453, 1454, 1455, 1457, 1462, 1469, 1473, 1474, 1476, 1477, 1478, 1481, 1487, 1488, 1489, 1490, 1496, 1499, 1500, 1502, 1503, 1504, 1507, 1508, 1510, 1512, 1516, 1517, 1518, 1525, 1528, 1533, 1535, 1536, 1542, 1545, 1546, 1547, 1549, 1551, 1552, 1558, 1560, 1561, 1562, 1568, 1574, 1581, 1585, 1589, 1590, 1591, 1593, 1594, 1600, 1602, 1603, 1604, 1605, 1606, 1607, 1608, 1611, 1612, 1613, 1619, 1622, 1624, 1627, 1631, 1632, 1633, 1637, 1638, 1639, 1641, 1647, 1648, 1650, 1654, 1655, 1656, 1657, 1658, 1659, 1660, 1668, 1669, 1671, 1672, 1674, 1677, 1678, 1679, 1680, 1684, 1686, 1691, 1694, 1695, 1697, 1698, 1701, 1710, 1711, 1713, 1716, 1717, 1718, 1720, 1721, 1722, 1726, 1729, 1733, 1734, 1735, 1736, 1737, 1738, 1739, 1740, 1743, 1745, 1747, 1751, 1753, 1754, 1755, 1756, 1758, 1759, 1765, 1766, 1767, 1770, 1771, 1778, 1779, 1780, 1783, 1784, 1785, 1786, 1787, 1788, 1791, 1794, 1798, 1799, 1800, 1805, 1809, 1810, 1812, 1816, 1818, 1824, 1825, 1827, 1833, 1834, 1835, 1840, 1841, 1848, 1849, 1852, 1853, 1855, 1856, 1857, 1859, 1860, 1862, 1866, 1867, 1869, 1870, 1877, 1879, 1882, 1888, 1889, 1890, 1892, 1897, 1900, 1901, 1904, 1909, 1915, 1919, 1920, 1923, 1926, 1928, 1930, 1931, 1933, 1934, 1935, 1936, 1938, 1940, 1941, 1944, 1948, 1949, 1951, 1953, 1955, 1958, 1959, 1961, 1962, 1963, 1964, 1965, 1970, 1971, 1973, 1974, 1978, 1979, 1981, 1984, 1988, 1989, 1990, 1992, 1994, 1995, 1996, 1998, 2000, 2001, 2002, 2008, 2013, 2014, 2017, 2019, 2020, 2025, 2035, 2036, 2038, 2039, 2041, 2042, 2046, 2048, 2049, 2052, 2054, 2058, 2063, 2064, 2070, 2072, 2077, 2078, 2079, 2081, 2082, 2085, 2087, 2088, 2091, 2092, 2095, 2099, 2100, 2102, 2104, 2106, 2109, 2110, 2112, 2114, 2119, 2121, 2124, 2126, 2131, 2132, 2133, 2135, 2138, 2139, 2141, 2142, 2145, 2146, 2147, 2156, 2159, 2160, 2161, 2162, 2164, 2167, 2169, 2173, 2177, 2180, 2181, 2183, 2184, 2185, 2186, 2187, 2188, 2192, 2193, 2194, 2195, 2196, 2198, 2199, 2200, 2203, 2204, 2212, 2215, 2219, 2222, 2223, 2226, 2230, 2231, 2233, 2240, 2241, 2242, 2243, 2245, 2247, 2248, 2250, 2253, 2256, 2258, 2260, 2261, 2262, 2263, 2265, 2267, 2274, 2278, 2280, 2285, 2287, 2289, 2290, 2297, 2298, 2301, 2304, 2306, 2311, 2313, 2317, 2319, 2322, 2324, 2333, 2335, 2337, 2338, 2339, 2342, 2344, 2348, 2353, 2360, 2361, 2362, 2364, 2366, 2367, 2372, 2376, 2378, 2379, 2382, 2389, 2390, 2395, 2400, 2404, 2405, 2406, 2408, 2409, 2410, 2412, 2416, 2420, 2421, 2423, 2424, 2425, 2426, 2437, 2440, 2441, 2442, 2444, 2445, 2446, 2448, 2450, 2453, 2454, 2459, 2461, 2463, 2464, 2465, 2470, 2472, 2473, 2475, 2477, 2481, 2488, 2491, 2493, 2496, 2497, 2499, 2500, 2501, 2507, 2509, 2510, 2511, 2512, 2513, 2515, 2517, 2518, 2520, 2522, 2526, 2529, 2530, 2532, 2535, 2537, 2538, 2539, 2540, 2544, 2547, 2548, 2549, 2551, 2553, 2556, 2558, 2559, 2562, 2570, 2574, 2575, 2576, 2577, 2578, 2586, 2589, 2590, 2592, 2593, 2594, 2595, 2602, 2605, 2606, 2607, 2609, 2610, 2616, 2620, 2622, 2623, 2624, 2625, 2626, 2628, 2629, 2630, 2631, 2633, 2635, 2636, 2638, 2641, 2644, 2645, 2647, 2653, 2663, 2664, 2665, 2667, 2672, 2676, 2678, 2685, 2687, 2688, 2694, 2701, 2702, 2703, 2707, 2708, 2713, 2714, 2716, 2717, 2722, 2726, 2727, 2729, 2730, 2731, 2732, 2733, 2734, 2738, 2742, 2744, 2746, 2752, 2754, 2756, 2757, 2759, 2760, 2761, 2762, 2764, 2765, 2766, 2767, 2768, 2769, 2771, 2772, 2774, 2776, 2777, 2778, 2781, 2788, 2791, 2792, 2793, 2795, 2798, 2800, 2801, 2803, 2810, 2811, 2813, 2816, 2820, 2821, 2824, 2829, 2836, 2837, 2841, 2843, 2845, 2846, 2850, 2851, 2854, 2857, 2858, 2860, 2862, 2863, 2864, 2867, 2870, 2873, 2875, 2877, 2878, 2879, 2880, 2881, 2882, 2884, 2889, 2892, 2897, 2899, 2902, 2905, 2910, 2912, 2915, 2916, 2921, 2922, 2925, 2929, 2930, 2933, 2935, 2937, 2938, 2939, 2942, 2944, 2946, 2949, 2951, 2953, 2954, 2955, 2956, 2960, 2963, 2964, 2967, 2968, 2971, 2972, 2975, 2976, 2978, 2979, 2980, 2981, 2982, 2983, 2985, 2991, 2994, 2995, 2996, 2998, 2999]}

logs/internlm2_5-7b-chat-1m_epoch_4.txt

@@ -1,157 +1,7 @@
 loading env vars from: /common2/dh.huang.2023/code/logical-reasoning/.env
 Adding /common2/dh.huang.2023/code/logical-reasoning to sys.path
 loading /common2/dh.huang.2023/code/logical-reasoning/llm_toolkit/logical_reasoning_utils.py
-internlm/internlm2_5-7b-chat-1m llama-factory/saves/internlm2_5_7b/lora/sft_bf16_p2_full/checkpoint-352 False datasets/mgtv results/mgtv-results_p2_full.csv
+internlm/internlm2_5-7b-chat-1m llama-factory/saves/internlm2_5_7b/lora/sft_bf16_p2_full_r2/checkpoint-700 False datasets/mgtv results/mgtv-results_p2_full_r2.csv
 (1) GPU = NVIDIA L40. Max memory = 44.309 GB.
 0.0 GB of memory reserved.
-loading model: internlm/internlm2_5-7b-chat-1m
-07/14/2024 11:08:55 - INFO - llamafactory.data.template - Replace eos token: <|im_end|>
-07/14/2024 11:08:55 - INFO - llamafactory.data.template - Add <|im_start|> to stop words.
-07/14/2024 11:08:55 - INFO - llamafactory.model.patcher - Using KV cache for faster generation.
-07/14/2024 11:09:06 - INFO - llamafactory.model.model_utils.attention - Using vanilla attention implementation.
-07/14/2024 11:09:06 - INFO - llamafactory.model.adapter - Merged 1 adapter(s).
-07/14/2024 11:09:06 - INFO - llamafactory.model.adapter - Loaded adapter(s): llama-factory/saves/internlm2_5_7b/lora/sft_bf16_p2_full/checkpoint-352
-07/14/2024 11:09:06 - INFO - llamafactory.model.loader - all params: 7,737,708,544
-(2) GPU = NVIDIA L40. Max memory = 44.309 GB.
-14.855 GB of memory reserved.
-loading train/test data files
-DatasetDict({
-    train: Dataset({
-        features: ['text', 'label', 'answer', 'title', 'puzzle', 'truth', 'train_text', 'prompt'],
-        num_rows: 25000
-    })
-    test: Dataset({
-        features: ['text', 'label', 'answer', 'title', 'puzzle', 'truth', 'train_text', 'prompt'],
-        num_rows: 3000
-    })
-})
---------------------------------------------------
-text: 甄加索是自杀吗
---------------------------------------------------
-label: 不是
---------------------------------------------------
-answer: nan
---------------------------------------------------
-title: 海岸之谜
---------------------------------------------------
-puzzle: 在远离城市喧嚣的海边小屋，一天清晨，邻居发现甄加索僵卧在沙滩上，已无生命迹象。现场没有发现任何打斗的迹象。请问甄加索的死因是什么？
---------------------------------------------------
-truth: 甄加索是一位热爱自然的画家，他每年都会来到这个海边小屋寻找灵感。在他生命的最后几天，他一直在创作一幅描绘海洋生物的画作。在画即将完成的前一天晚上，他骑着自行车外出，打算在海边观赏夜景。然而，他在沙滩上意外发现了一只搁浅的海豚，为了救助这只海豚，他耗费了极大的体力，最终成功将其送回海中。筋疲力尽的甄加索在沙滩上睡着了，由于他患有严重的心脏病，却未告知旁人，在寒冷的海风中，他的心脏停止了跳动。因此，警方在现场只发现了车轮痕迹和未完成的画作，而没有发现任何他杀的迹象。
---------------------------------------------------
-train_text: <|im_start|>system
-You are an expert in logical reasoning.<|im_end|>
-<|im_start|>user
-你是一个逻辑游戏的主持人。游戏规则如下：
-
-1. 参与者会得到一个谜题。
-2. 参与者可以通过提问来获取线索，尝试解开谜题。
-3. 对于每个问题，主持人将根据实际情况回答以下五个选项之一：是、不是、不重要、回答正确、问法错误。
-4. 回答中不能添加任何其它信息，也不能省略选项中的任何一个字。例如，不可以把“不是”省略成“不”。
-5. 参与者需要根据回答来推理，并最终找出谜题的正确答案。
-
-请严格按照这些规则回答参与者提出的问题。
-
-谜题: 在远离城市喧嚣的海边小屋，一天清晨，邻居发现甄加索僵卧在沙滩上，已无生命迹象。现场没有发现任何打斗的迹象。请问甄加索的死因是什么？
-
-实际情况: 甄加索是一位热爱自然的画家，他每年都会来到这个海边小屋寻找灵感。在他生命的最后几天，他一直在创作一幅描绘海洋生物的画作。在画即将完成的前一天晚上，他骑着自行车外出，打算在海边观赏夜景。然而，他在沙滩上意外发现了一只搁浅的海豚，为了救助这只海豚，他耗费了极大的体力，最终成功将其送回海中。筋疲力尽的甄加索在沙滩上睡着了，由于他患有严重的心脏病，却未告知旁人，在寒冷的海风中，他的心脏停止了跳动。因此，警方在现场只发现了车轮痕迹和未完成的画作，而没有发现任何他杀的迹象。
-
-参与者提出的问题: 甄加索是自杀吗
-<|im_end|>
-<|im_start|>assistant
-不是<|im_end|>
---------------------------------------------------
-prompt: <|im_start|>system
-You are an expert in logical reasoning.<|im_end|>
-<|im_start|>user
-你是一个逻辑游戏的主持人。游戏规则如下：
-
-1. 参与者会得到一个谜题。
-2. 参与者可以通过提问来获取线索，尝试解开谜题。
-3. 对于每个问题，主持人将根据实际情况回答以下五个选项之一：是、不是、不重要、回答正确、问法错误。
-4. 回答中不能添加任何其它信息，也不能省略选项中的任何一个字。例如，不可以把“不是”省略成“不”。
-5. 参与者需要根据回答来推理，并最终找出谜题的正确答案。
-
-请严格按照这些规则回答参与者提出的问题。
-
-谜题: 在远离城市喧嚣的海边小屋，一天清晨，邻居发现甄加索僵卧在沙滩上，已无生命迹象。现场没有发现任何打斗的迹象。请问甄加索的死因是什么？
-
-实际情况: 甄加索是一位热爱自然的画家，他每年都会来到这个海边小屋寻找灵感。在他生命的最后几天，他一直在创作一幅描绘海洋生物的画作。在画即将完成的前一天晚上，他骑着自行车外出，打算在海边观赏夜景。然而，他在沙滩上意外发现了一只搁浅的海豚，为了救助这只海豚，他耗费了极大的体力，最终成功将其送回海中。筋疲力尽的甄加索在沙滩上睡着了，由于他患有严重的心脏病，却未告知旁人，在寒冷的海风中，他的心脏停止了跳动。因此，警方在现场只发现了车轮痕迹和未完成的画作，而没有发现任何他杀的迹象。
-
-参与者提出的问题: 甄加索是自杀吗
-<|im_end|>
-<|im_start|>assistant
-
---------------------------------------------------
-text: 死者受伤了吗
---------------------------------------------------
-label: 不是
---------------------------------------------------
-answer: nan
---------------------------------------------------
-title: 甄庄哭声
---------------------------------------------------
-puzzle: 在一个安静的夜晚，小村庄的湖边突然传来了阵阵哭泣声。第二天早晨，村长甄锐发现湖边的石头上放着一顶破旧的帽子，但没有人知道这顶帽子是从哪里来的，哭泣声又是为何。请还原故事真相。
---------------------------------------------------
-truth: 原来，这顶破旧的帽子属于一个小男孩，他小时候与爷爷在湖边生活。爷爷教他钓鱼、游泳，还告诉他湖中的海龟是他们的朋友。后来，小男孩随父母去了城市生活，但每年夏天都会回到村子探望爷爷。然而，去年夏天，爷爷因病去世，小男孩伤心欲绝。今年夏天，他回到村子，来到湖边，想起和爷爷的美好回忆，忍不住哭泣。他将爷爷的帽子放在湖边的石头上，希望能让爷爷的在天之灵得到安慰。那晚的哭泣声正是小男孩在祭莫他亲爱的爷爷。
---------------------------------------------------
-train_text: <|im_start|>system
-You are an expert in logical reasoning.<|im_end|>
-<|im_start|>user
-你是一个逻辑游戏的主持人。游戏规则如下：
-
-1. 参与者会得到一个谜题。
-2. 参与者可以通过提问来获取线索，尝试解开谜题。
-3. 对于每个问题，主持人将根据实际情况回答以下五个选项之一：是、不是、不重要、回答正确、问法错误。
-4. 回答中不能添加任何其它信息，也不能省略选项中的任何一个字。例如，不可以把“不是”省略成“不”。
-5. 参与者需要根据回答来推理，并最终找出谜题的正确答案。
-
-请严格按照这些规则回答参与者提出的问题。
-
-谜题: 在一个安静的夜晚，小村庄的湖边突然传来了阵阵哭泣声。第二天早晨，村长甄锐发现湖边的石头上放着一顶破旧的帽子，但没有人知道这顶帽子是从哪里来的，哭泣声又是为何。请还原故事真相。
-
-实际情况: 原来，这顶破旧的帽子属于一个小男孩，他小时候与爷爷在湖边生活。爷爷教他钓鱼、游泳，还告诉他湖中的海龟是他们的朋友。后来，小男孩随父母去了城市生活，但每年夏天都会回到村子探望爷爷。然而，去年夏天，爷爷因病去世，小男孩伤心欲绝。今年夏天，他回到村子，来到湖边，想起和爷爷的美好回忆，忍不住哭泣。他将爷爷的帽子放在湖边的石头上，希望能让爷爷的在天之灵得到安慰。那晚的哭泣声正是小男孩在祭莫他亲爱的爷爷。
-
-参与者提出的问题: 死者受伤了吗
-<|im_end|>
-<|im_start|>assistant
-不是<|im_end|>
---------------------------------------------------
-prompt: <|im_start|>system
-You are an expert in logical reasoning.<|im_end|>
-<|im_start|>user
-你是一个逻辑游戏的主持人。游戏规则如下：
-
-1. 参与者会得到一个谜题。
-2. 参与者可以通过提问来获取线索，尝试解开谜题。
-3. 对于每个问题，主持人将根据实际情况回答以下五个选项之一：是、不是、不重要、回答正确、问法错误。
-4. 回答中不能添加任何其它信息，也不���省略选项中的任何一个字。例如，不可以把“不是”省略成“不”。
-5. 参与者需要根据回答来推理，并最终找出谜题的正确答案。
-
-请严格按照这些规则回答参与者提出的问题。
-
-谜题: 在一个安静的夜晚，小村庄的湖边突然传来了阵阵哭泣声。第二天早晨，村长甄锐发现湖边的石头上放着一顶破旧的帽子，但没有人知道这顶帽子是从哪里来的，哭泣声又是为何。请还原故事真相。
-
-实际情况: 原来，这顶破旧的帽子属于一个小男孩，他小时候与爷爷在湖边生活。爷爷教他钓鱼、游泳，还告诉他湖中的海龟是他们的朋友。后来，小男孩随父母去了城市生活，但每年夏天都会回到村子探望爷爷。然而，去年夏天，爷爷因病去世，小男孩伤心欲绝。今年夏天，他回到村子，来到湖边，想起和爷爷的美好回忆，忍不住哭泣。他将爷爷的帽子放在湖边的石头上，希望能让爷爷的在天之灵得到安慰。那晚的哭泣声正是小男孩在祭莫他亲爱的爷爷。
-
-参与者提出的问题: 死者受伤了吗
-<|im_end|>
-<|im_start|>assistant
-
-Evaluating model: internlm/internlm2_5-7b-chat-1m
---------
-step 1: 不是</s>
---------
-step 2: 不是
---------
-step 3: 不是
---------
-step 4: 不是
---------
-step 5: 不是
-(3) GPU = NVIDIA L40. Max memory = 44.309 GB.
-16.078 GB of memory reserved.
-      text  ... internlm/internlm2_5-7b-chat-1m_checkpoint-352
-0  甄加索是自杀吗  ...                                             不是
-
-[1 rows x 10 columns]
-{'accuracy': 0.7303333333333333, 'incorrect_ids': [9, 13, 15, 16, 24, 25, 29, 31, 33, 34, 35, 36, 37, 38, 55, 57, 58, 59, 61, 65, 66, 67, 78, 82, 83, 84, 88, 91, 93, 94, 96, 97, 99, 104, 106, 108, 109, 112, 119, 121, 124, 128, 129, 135, 137, 138, 139, 141, 143, 150, 153, 155, 164, 171, 172, 173, 179, 190, 191, 192, 199, 200, 202, 207, 218, 222, 224, 225, 226, 227, 228, 230, 234, 236, 245, 248, 250, 251, 252, 253, 255, 257, 259, 260, 265, 269, 271, 272, 273, 275, 276, 283, 284, 286, 292, 299, 301, 303, 304, 309, 311, 314, 316, 318, 321, 323, 325, 334, 335, 337, 338, 339, 341, 342, 347, 350, 352, 353, 354, 355, 356, 357, 360, 368, 370, 371, 373, 374, 377, 383, 386, 389, 395, 397, 404, 410, 421, 423, 428, 429, 430, 432, 438, 440, 445, 447, 452, 454, 456, 457, 469, 470, 471, 472, 473, 476, 480, 482, 490, 493, 494, 495, 497, 500, 501, 502, 506, 507, 508, 510, 511, 514, 517, 518, 520, 530, 533, 536, 537, 538, 540, 553, 560, 566, 570, 571, 579, 581, 585, 589, 591, 592, 593, 596, 597, 599, 601, 612, 613, 614, 621, 625, 626, 627, 628, 632, 636, 644, 647, 650, 657, 663, 666, 671, 678, 682, 686, 693, 695, 701, 702, 708, 709, 718, 720, 721, 722, 726, 727, 729, 730, 732, 734, 739, 740, 752, 754, 758, 760, 766, 770, 771, 772, 774, 778, 788, 790, 791, 794, 795, 798, 800, 801, 803, 805, 809, 817, 819, 820, 821, 823, 824, 828, 833, 837, 840, 842, 843, 844, 847, 851, 856, 857, 861, 862, 864, 866, 869, 870, 875, 884, 888, 889, 890, 899, 901, 904, 913, 927, 935, 937, 943, 945, 952, 958, 962, 966, 969, 970, 980, 982, 989, 991, 994, 997, 998, 1001, 1003, 1005, 1006, 1007, 1011, 1012, 1013, 1014, 1015, 1022, 1036, 1045, 1046, 1051, 1053, 1061, 1066, 1071, 1075, 1076, 1080, 1081, 1083, 1087, 1089, 1096, 1111, 1116, 1120, 1121, 1125, 1126, 1129, 1135, 1138, 1154, 1158, 1163, 1164, 1166, 1167, 1170, 1174, 1177, 1178, 1179, 1180, 1181, 1183, 1185, 1192, 1194, 1203, 1212, 1219, 1228, 1232, 1239, 1240, 1241, 1242, 1251, 1252, 1254, 1259, 1266, 1282, 1289, 1293, 1300, 1301, 1305, 1308, 1311, 1313, 1314, 1315, 1317, 1323, 1324, 1332, 1339, 1342, 1347, 1349, 1353, 1357, 1363, 1364, 1367, 1369, 1370, 1373, 1380, 1385, 1387, 1388, 1389, 1392, 1393, 1400, 1406, 1407, 1409, 1412, 1413, 1420, 1426, 1427, 1438, 1440, 1444, 1445, 1448, 1451, 1453, 1454, 1455, 1459, 1460, 1462, 1469, 1470, 1473, 1475, 1476, 1478, 1481, 1486, 1491, 1493, 1494, 1496, 1500, 1501, 1512, 1515, 1517, 1518, 1525, 1526, 1528, 1533, 1544, 1547, 1554, 1559, 1560, 1562, 1567, 1573, 1581, 1584, 1585, 1586, 1590, 1593, 1596, 1603, 1604, 1605, 1606, 1611, 1613, 1622, 1627, 1633, 1637, 1641, 1645, 1647, 1648, 1650, 1654, 1655, 1658, 1659, 1660, 1665, 1667, 1668, 1672, 1673, 1674, 1676, 1679, 1683, 1686, 1689, 1691, 1695, 1700, 1716, 1718, 1726, 1727, 1730, 1734, 1736, 1738, 1746, 1751, 1752, 1755, 1756, 1769, 1770, 1773, 1780, 1786, 1790, 1791, 1796, 1799, 1800, 1801, 1804, 1809, 1812, 1816, 1825, 1827, 1833, 1835, 1848, 1854, 1858, 1860, 1867, 1869, 1879, 1881, 1888, 1897, 1903, 1907, 1913, 1914, 1933, 1935, 1940, 1944, 1945, 1949, 1953, 1958, 1963, 1964, 1965, 1970, 1973, 1978, 1981, 1984, 1989, 1990, 1991, 1992, 1994, 1995, 1998, 2001, 2005, 2007, 2010, 2014, 2015, 2017, 2021, 2029, 2035, 2036, 2037, 2038, 2040, 2046, 2049, 2062, 2063, 2067, 2072, 2076, 2077, 2094, 2106, 2109, 2110, 2112, 2118, 2119, 2121, 2122, 2126, 2130, 2133, 2135, 2140, 2141, 2145, 2147, 2161, 2162, 2164, 2167, 2169, 2177, 2185, 2186, 2189, 2193, 2194, 2195, 2197, 2199, 2208, 2212, 2222, 2226, 2230, 2233, 2237, 2240, 2246, 2248, 2249, 2250, 2254, 2257, 2261, 2262, 2265, 2274, 2275, 2281, 2285, 2297, 2304, 2309, 2311, 2313, 2319, 2320, 2322, 2324, 2333, 2335, 2344, 2348, 2354, 2360, 2364, 2365, 2366, 2378, 2386, 2389, 2395, 2396, 2400, 2404, 2406, 2408, 2409, 2410, 2412, 2421, 2423, 2425, 2429, 2437, 2440, 2441, 2442, 2444, 2448, 2463, 2469, 2471, 2484, 2486, 2491, 2493, 2512, 2517, 2522, 2526, 2529, 2530, 2532, 2535, 2538, 2539, 2546, 2547, 2548, 2549, 2553, 2554, 2555, 2556, 2559, 2560, 2562, 2567, 2574, 2575, 2581, 2589, 2590, 2594, 2602, 2607, 2616, 2624, 2626, 2629, 2632, 2635, 2644, 2647, 2655, 2660, 2663, 2667, 2672, 2676, 2681, 2691, 2704, 2707, 2710, 2714, 2715, 2716, 2727, 2728, 2731, 2736, 2742, 2744, 2745, 2749, 2754, 2756, 2757, 2760, 2761, 2762, 2764, 2766, 2767, 2777, 2788, 2797, 2798, 2801, 2803, 2806, 2807, 2811, 2812, 2815, 2816, 2820, 2823, 2837, 2843, 2845, 2850, 2854, 2857, 2877, 2882, 2884, 2888, 2899, 2902, 2905, 2912, 2913, 2915, 2916, 2921, 2927, 2933, 2937, 2938, 2942, 2944, 2949, 2953, 2960, 2962, 2969, 2973, 2975, 2979, 2983, 2991, 2995, 2998, 2999]}
+loading model: internlm/internlm2_5-7b-chat-1m with adapter: llama-factory/saves/internlm2_5_7b/lora/sft_bf16_p2_full_r2/checkpoint-700

scripts/eval-mgtv-glm-4-9b.sh

@@ -19,6 +19,8 @@ pip install -r requirements.txt
 export LOGICAL_REASONING_DATA_PATH=datasets/mgtv
 export RESIZE_TOKEN_EMBEDDINGS=true
 export START_EPOCH=1
+export USING_LLAMA_FACTORY=false
+
 
 export MODEL_NAME=THUDM/glm-4-9b-chat-1m
 export MODEL_PREFIX=glm-4-9b_lora_sft_bf16
@@ -26,10 +28,9 @@ export MODEL_PREFIX=glm-4-9b_lora_sft_bf16
 export LOGICAL_REASONING_RESULTS_PATH=results/$MODEL_PREFIX-p1.csv
 export ADAPTER_PATH_BASE=llama-factory/saves/glm-4-9b/lora/sft_bf16_p1_full
 export USING_P1_PROMPT_TEMPLATE=true
-#echo "Eval $MODEL_NAME with $ADAPTER_PATH_BASE"
-#python llm_toolkit/eval_logical_reasoning_all_epochs.py
+echo "Eval $MODEL_NAME with $ADAPTER_PATH_BASE"
+python llm_toolkit/eval_logical_reasoning_all_epochs.py
 
-export START_EPOCH=4
 export LOGICAL_REASONING_RESULTS_PATH=results/$MODEL_PREFIX-p2.csv
 export ADAPTER_PATH_BASE=llama-factory/saves/glm-4-9b/lora/sft_bf16_p2_full
 export USING_P1_PROMPT_TEMPLATE=false

scripts/eval-mgtv-internlm.sh

@@ -21,6 +21,7 @@ export LOGICAL_REASONING_DATA_PATH=datasets/mgtv
 export LOGICAL_REASONING_RESULTS_PATH=results/mgtv-results_p2_full_r2.csv
 export ADAPTER_PATH_BASE=llama-factory/saves/internlm2_5_7b/lora/sft_bf16_p2_full_r2
 export USING_LLAMA_FACTORY=true
+export START_EPOCH=4
 
 echo "Eval $MODEL_NAME with $ADAPTER_PATH_BASE"
 python llm_toolkit/eval_logical_reasoning_all_epochs.py