trainer_bpe_v2 / config.json
onnookk's picture
Training in progress, step 1500
4288170 verified
{
"_name_or_path": "axiong/PMC_LLaMA_13B",
"architectures": [
"LlamaForCausalLM_bpe"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 0,
"eos_token_id": 1,
"hidden_act": "silu",
"hidden_size": 5120,
"initializer_range": 0.02,
"intermediate_size": 13824,
"mask": [
13,
2277,
2277,
29937,
29889,
13,
13,
13,
29871,
29896,
29900,
29889,
29871,
29906,
3030,
29889,
29901,
13,
29900,
29900,
29889,
29900,
310,
278,
29915,
29879,
29889,
450,
297,
278,
29871,
29900,
29892,
29871,
411,
385,
29892,
322,
278,
16500,
338,
385,
29871,
29941,
16500,
29915,
278,
1139,
278,
16083,
3300,
2859,
322,
29871,
263,
2933,
7128,
2486,
10608,
29889,
3414,
29892,
385,
15278,
2009,
29889,
29914,
1217,
13866,
338,
15278,
393,
393,
16612,
16612,
263,
263,
3414,
29892,
3300,
2859,
411,
385,
1881,
1881,
393,
393,
8128,
8128,
4340,
4340,
3030,
29889,
6113,
6113,
263,
2933,
393,
393,
7128,
2486,
1614,
1614,
2167,
2167,
278,
278,
2009,
29937,
2799,
2799,
4080,
4080,
29901,
13,
2744,
2744,
14997,
14997,
911,
911,
278,
1139,
2183,
2183,
967,
967,
3030,
29889,
25538,
25538,
1716,
1716,
1472,
1472,
1234,
1234,
322,
322,
4874,
4874,
29914,
1217,
10608,
29937,
10567,
10567,
29901,
29937,
16492,
16492,
29901,
29937,
2677,
2677,
29901,
29937,
13291,
13291,
29901,
29973,
13,
16083,
5155,
29906,
29900,
29896,
29889,
29900,
29896,
29879,
6139,
310,
29871,
353,
29871,
29871,
29945,
263,
11619,
29871,
29946,
29896,
29900,
304,
278,
29871,
29929,
11619,
29892,
408,
263,
6139,
29889,
22069,
411,
29871,
29953,
29945,
29995,
29897,
322,
373,
278,
29889,
29945,
29906,
29889,
297,
29871,
29929,
29945,
29889,
29906,
29889,
29941,
29889,
29896,
304,
29871,
29889,
29946,
29871,
29955,
29871,
29947,
29889,
29955,
29896,
29929,
313,
29925,
322,
278,
29889,
29953,
313,
29896,
29896,
29906,
29892,
3113,
29892,
278,
1234,
278,
29889,
29929,
29966,
29900,
313,
29886,
29941,
29889,
29889,
319,
29889,
29947,
471,
29871,
3372,
1230,
29889,
512,
29889,
1334,
29896,
29896,
29900,
29945,
29896,
29945,
29946,
29889,
2729,
373,
304,
1234,
411,
263,
25300,
708,
29896,
29941,
6942,
411,
24899,
936,
29945,
29889,
29941,
29900,
313,
29906,
29896,
467,
29947,
29889,
29899,
29896,
29929,
29929,
29953,
29889,
29955,
29889,
29995,
310,
29995,
25781,
467,
13,
29900,
29995,
773,
278,
338,
304,
29896,
29946,
29896,
29947,
6559,
471,
363,
278,
313,
29876,
411,
278,
29906,
29945,
29889,
29871,
29896,
29897,
29901,
29871,
10503,
2561,
313,
29941,
467,
450,
313,
29929,
5155,
2729,
5758,
2877,
29906,
29946,
29901,
1763,
278,
29871,
22069,
313,
29945,
29900,
29896,
29892,
29892,
596,
29899,
786,
310,
22069,
515,
29871,
29929,
29889,
3633,
29889,
445,
6559,
29892,
25036,
29879,
3633,
297,
263,
310,
445,
471,
304,
29900,
29906,
29900,
29899,
5155,
29889,
3211,
278,
29896,
29953,
511,
322,
1101,
29899,
29901,
5538,
1546,
278,
13,
29954,
29954,
5428,
5428,
596,
3113,
3867,
1661,
29899,
338,
263,
263,
16083,
29901,
1317,
29906,
29897,
29901,
450,
5034,
304,
13,
2887,
2887,
263,
10257,
29892,
29896,
29955,
29936,
29871,
297,
22069,
5155,
773,
9401,
411,
29906,
29896,
6297,
408,
24876,
19263,
29897,
892,
29893,
296,
29896,
29899,
1090,
29893,
13,
10858,
758,
5138,
7186,
29889,
310,
263,
14707,
278,
29922,
29900,
29889,
910,
29899,
29906,
29946,
29900,
29900,
29941,
29953,
29900,
472,
29871,
29906,
29899,
1400,
3372,
1353,
310,
22069,
892,
515,
278,
22069,
29889,
313,
29946,
4121,
10070,
29897,
297,
29906,
29906,
29995,
313,
29906,
29941,
3240,
1883,
1223,
11517,
29220,
27580,
714,
26807,
304,
8161,
29892,
8307,
29945,
29899,
25300,
936,
292,
278,
29900,
29929,
29941,
29953,
25828,
4835,
29995,
511,
29876,
353,
29941,
29945,
671,
310,
1258,
451,
29941,
29995,
29900,
29946,
29906,
29953,
718,
24028,
29889,
4121,
29955,
29995,
472,
278,
29941,
29941,
29929,
29900,
29945,
467,
29889,
1670,
29906,
29892,
313,
29953,
29886,
29966,
263,
29871,
292,
263,
359,
9708,
29892,
349,
29941,
29896,
29946,
29995,
24028,
29871,
2398,
29892,
29899,
6360,
29941,
29906,
724,
398,
30080,
29900,
304,
367,
22069,
29892,
313,
1955,
29900,
29947,
4120,
654,
3483,
952,
16165,
441,
29906,
29947,
9401,
304,
29892,
411,
29900,
29953,
29871,
30221,
278,
6559,
313,
29945,
12997,
1711,
30221,
29871,
3629,
292,
29900,
29955,
1883,
12645,
29929,
29947,
29946,
29945,
29955,
29945,
29941,
29946,
30080,
29922,
29922,
30080,
22069,
1058,
29906,
29995,
29925,
353,
29886,
353,
29947,
29900,
29892,
372,
13,
3629,
596,
3414,
3414,
338,
27432,
403,
750,
263,
285,
1461,
1304,
304,
29889,
2178,
29896,
29995,
491,
278,
9045,
18020,
892,
29871,
938,
336,
6559,
29889,
29900,
29892,
29925,
29966,
1274,
1082,
654,
261,
372,
338,
2318,
29889,
29892,
282,
752,
5795,
6151,
1070,
29906,
29929,
9708,
293,
3001,
310,
263,
441,
25781,
29871,
289,
650,
29897,
471,
29945,
29892,
471,
8560,
29953,
29945,
2362,
5570,
29945,
29897,
29945,
2440,
29953,
29955,
652,
370,
29955,
29900,
5138,
6906,
29941,
29892,
29906,
29955,
29953,
29995,
322,
263,
363,
29871,
29941,
29955,
29537,
287,
1228,
27580,
29892,
263,
910,
6559,
313,
29955,
263,
9045,
29889,
2398,
13,
3492,
3492,
29915,
29915,
276,
276,
263,
25036,
3211,
16083,
9365,
9365,
5034,
6554,
310,
1189,
293,
29929,
29953,
7186,
29871,
29899,
29941,
29896,
511,
671,
278,
20890,
304,
29946,
29946,
17924,
895,
29946,
29906,
3211,
292,
441,
293,
23797,
29871,
1711,
7282,
16083,
4120,
261,
29892,
29892,
591,
410,
5138,
471,
1304,
758,
791,
895,
297,
19263,
310,
13,
13696,
13696,
1241,
1241,
292,
292,
596,
596,
6297,
3113,
671,
6139,
304,
29941,
29929,
5046,
29892,
450,
12242,
19220,
310,
297,
3211,
12242,
310,
282,
29966,
29892,
541,
25483,
297,
6471,
29889,
7163,
2200,
12045,
310,
278,
937,
29947,
29995,
10858,
10110,
10110,
338,
25036,
3867,
3867,
6089,
6089,
304,
5155,
411,
411,
19220,
18020,
10257,
3113,
14707,
6139,
322,
322,
5957,
5957,
596,
596,
17924,
297,
22862,
22862,
278,
596,
3239,
3239,
408,
3867,
596,
596,
25483,
1135,
29871,
29995,
7186,
29946,
29899,
29946,
29947,
29945,
29946,
24809,
358,
791,
663,
29945,
29941,
29929,
29995,
2761,
2318,
2318,
313,
596,
6351,
6351,
408,
3867,
20890,
411,
29871,
29945,
29953,
349,
29966,
29929,
29906,
370,
10778,
410,
3859,
304,
3211,
29946,
29941,
29995,
297,
29941,
29947,
282,
353,
758,
3372,
29929,
29955,
313,
29900,
29941,
29899,
17168,
293,
29946,
29892,
349,
353,
1559,
16381,
29946,
29896,
29906,
467,
892,
9401,
952,
267,
319,
3001,
313,
29947,
278,
22069,
1539,
579,
29941,
29897,
29945,
29955,
29929,
29946,
29946,
29929,
966,
1080,
16083,
10257,
596,
23134,
23134,
338,
29899,
12817,
471,
263,
964,
3633,
1670,
892,
4392,
1312,
13284,
408,
338,
3806,
337,
5890,
5890,
373,
29901,
1815,
304,
263,
29953,
29906,
425,
862,
29995,
322,
8307,
29889,
471,
451,
5147,
29894,
29894,
936,
29945,
29945,
13,
797,
797,
596,
596,
13284,
3806,
393,
393,
366,
366,
1234,
5155,
337,
471,
16951,
353,
869,
29899,
29953,
322,
297,
29945,
29947,
450,
2099,
29947,
29906,
29889,
1939,
512,
278,
4768,
3554,
29946,
29953,
286,
29887,
4475,
304,
5528,
5084,
29995,
16420,
29953,
29899,
304,
14707,
16381,
4125,
29929,
29896,
24207,
23900,
7498,
10700,
5622,
964,
10122,
310,
29892,
470,
29899,
6707,
1546,
29871,
29898,
29906,
29901,
1938,
3633,
278,
29889,
4587,
29966,
29889,
22069,
1090,
297,
20309,
6382,
292,
278,
1021,
29900,
2440,
16420,
7292,
4328,
297,
437,
344,
287,
304,
312,
16103,
2779,
310,
7614,
293,
2775,
537,
9443,
1546,
24876,
2662,
22069,
750,
17905,
29872,
10791,
12645,
471,
694,
1773,
27432,
21622,
272,
522,
16103,
29892,
297,
287,
29889,
9076,
287,
1583,
29899,
2645,
278,
359,
8552,
29946,
29955,
29896,
322,
10858,
6297,
11619,
6858,
6858,
366,
366,
304,
5155,
5622,
277,
1338,
29897,
310,
29947,
29892,
8161,
278,
29947,
29896,
322,
304,
29899,
29900,
11029,
310,
29897,
338,
756,
1063,
29946,
29897,
12651,
297,
16951,
6133,
29955,
29896,
29901,
338,
29947,
29947,
29882,
655,
30081,
29922,
16157,
29879,
5881,
29875,
304,
24809,
1095,
359,
29945,
29906,
29929,
29892,
1670,
471,
29897,
470,
29936,
349,
352,
262,
29879,
29889,
4315,
292,
297,
445,
663,
310,
25082,
653,
8161,
3692,
869,
29900,
362,
29889,
5138,
520,
520,
293,
7243,
1037,
22898,
4871,
26552,
29871,
29418,
277,
10678,
618,
29889,
1763,
313,
29934,
29955,
29929,
4392,
3381,
620,
9739,
1006,
7316,
1058,
892,
1284,
886,
327,
333,
29887,
29914,
766,
23367,
8717,
29882,
29955,
29947,
862,
359,
29906,
313,
29953,
29946,
29953,
29892,
29953,
29896,
13667,
282,
1461,
1973,
892,
5134,
1008,
284,
3041,
1934,
1058,
1090,
29953,
29947,
586,
6151,
29955,
29941,
892,
451,
471,
1476,
9045,
29891,
487,
312,
29922,
30081,
29955,
29953,
467,
512,
5517,
304,
29929,
29941,
29897,
363,
564,
357,
1139,
15421,
29955,
29946,
29945,
22069,
694,
7282,
9263,
1463,
29947,
29945,
14914,
411,
29901,
4683,
4036,
1891,
29945,
29896,
29955,
29906,
297,
599,
6437,
310,
29947,
29946,
310,
2834,
8855,
630,
25781,
29901,
450,
6437,
511,
29871,
4955,
310,
9238,
10678,
29945,
10997,
29900,
29897,
1090,
17696,
29897,
411,
7282,
4328,
5490,
29871,
29953,
29953,
10997,
322,
11176,
14703,
27231,
5391,
362,
310,
29875,
586,
12045,
13879,
316,
2590,
459,
493,
29900,
313,
538,
616,
324,
293,
7418,
310,
29899,
2042,
29947,
29929,
4771,
24858,
892,
1304,
278,
1023,
312,
284,
1156,
278,
29953,
467,
758,
29899,
29947,
29897,
29896,
313,
29879,
29897,
29953,
29929,
29945,
313,
29947,
29953,
29947,
29941,
29889,
1152,
272,
353,
2200,
1070,
310,
1438,
29899,
8489,
2318,
29892,
29955,
29955,
537,
310,
472,
3203,
29946,
467,
29899,
29946,
278,
2761,
363,
263,
408,
278,
542,
538,
3554,
29891,
29899,
29945,
29900,
286,
892,
1476,
29941,
313,
24899,
293,
260,
15118,
29941,
467,
962,
561,
358,
310,
29900,
22069,
892,
8560,
7405,
630,
393,
278,
5084,
310,
25828,
290,
278,
758,
3677,
747,
1023,
6471,
5866,
411,
29995,
363,
564,
22005,
278,
10122,
30080,
30221,
30221,
30080,
2440,
29897,
1020,
10859,
271,
358,
322,
1661,
5198,
348,
304,
23033,
313,
3881,
1058,
750,
3692,
278,
29925,
29922,
1400,
29899,
2440,
29889,
892,
16951,
263,
7282,
403,
7418,
4891,
29899,
2042,
284,
7282,
12651,
1162,
26841,
8500,
573,
708,
29889,
278,
2779,
5134,
297,
313,
29878,
301,
962,
29906,
322,
3372,
6703,
363,
22069,
14502,
310,
29899,
29929,
29879,
29892,
29906,
10997,
29891,
29899,
29955,
29892,
29892,
3704,
590,
542,
2440,
29892,
29945,
29929,
29947,
29899,
4695,
17855,
29906,
22069,
10416,
12959,
23033,
278,
652,
21780,
29947,
467,
14060,
545,
1480,
4695,
12645,
6559,
278,
9443,
29955,
29899,
491,
29871,
7418,
29889,
29929,
29897,
29945,
304,
4824,
8910,
30081,
29900,
338,
14969,
29947,
29955,
29941,
10997,
24899,
1711,
326,
575,
29879,
313,
1763,
8161,
278,
671,
9712,
2200,
29871,
30386,
379,
5667,
297,
1716,
7910,
297,
29889,
2860,
29889,
3118,
29892,
607,
12003,
2264,
594,
6737,
29899,
29882,
29888,
3958,
619,
369,
22069,
322,
304,
505,
1560,
17223,
599,
22069,
427,
24476,
29879,
310,
773,
263,
22069,
297,
766,
2098,
350,
10403,
29953,
313,
6559,
29892,
29892,
892,
278,
1556,
1880,
29899,
29953,
22069,
2440,
313,
310,
599,
594,
3901,
10997,
750,
505,
1063,
29889,
3630,
29929,
467,
285,
26310,
8950,
1228,
1883,
1103,
1103,
3598,
29879,
892,
5046,
29871,
310,
349,
313,
29907,
408,
1532,
29897,
313,
2629,
29871,
15477,
1546,
304,
7252,
29929,
22069,
29899,
322,
19257,
310,
29953,
29941,
11174,
310,
1906,
411,
3129,
979,
278,
14502,
17135,
313,
6559,
310,
29876,
29922,
322,
1400,
29899,
29947,
4158,
2380,
892,
15659,
29929,
29899,
29953,
29897,
330,
7614,
24809,
278,
29955,
29897,
29929,
313,
14953,
800,
14385,
310,
471,
278,
9045,
2562,
29896,
22069,
29892,
408,
29892,
727,
29900,
322,
29974,
24028,
287,
363,
892,
694,
4587,
278,
1135,
297,
29906,
511,
1316,
408,
29889,
315,
1775,
17196,
1682,
852,
491,
263,
29946,
22069,
5031,
993,
3144,
1682,
29995,
23797,
29947,
313,
1663,
352,
21622,
943,
3138,
29892,
29909,
29897,
515,
263,
313,
29940,
29889,
530,
995,
310,
29900,
10997,
29899,
14940,
262,
29899,
3620,
297,
471,
8900,
362,
322,
2861,
304,
310,
385,
1472,
29899,
29947,
10997,
292,
363,
17135,
29889,
6133,
297,
10997,
892,
29955,
10997,
29945,
511,
4344,
411,
727,
471,
1461,
545,
630,
411,
25781,
29892,
594,
264,
29946,
10997,
29941,
22069
],
"max_position_embeddings": 2048,
"max_sequence_length": 2048,
"model_type": "llama",
"num_attention_heads": 40,
"num_hidden_layers": 40,
"num_key_value_heads": 40,
"pad_token_id": -1,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.37.0.dev0",
"use_cache": true,
"vocab_size": 32001
}