1588 Commits

Author SHA1 Message Date
hiyouga
bf5ad34196 support distributed quantized training
Former-commit-id: 74ff23a4f36f859f791f7b4be6f1877edc68f12f
2023-06-06 17:39:41 +08:00
hiyouga
ac6f50dedf add API demo from #1
Former-commit-id: c955edcef168da44257c5b50d7bc59266d909782
2023-06-05 21:32:18 +08:00
hoshi-hiyouga
8fd9ef924d Merge pull request #11 from hiyouga/api
Api

Former-commit-id: 9b2f524ea7f3a28f7413b8ce67e585f1596566a5
2023-06-05 20:58:02 +08:00
hiyouga
a409e1f42c fix bug in web demo
Former-commit-id: 01d6d7a910b9845a0ea38632661ce813e5cfe3a2
2023-06-05 17:58:29 +08:00
hiyouga
3f5869111b increase max length in cli demo
Former-commit-id: 0113cdb12728419022b5c01c932a5d52e626a200
2023-06-05 16:49:14 +08:00
hiyouga
f9c51a8340 implement stream generating
Former-commit-id: 6cc9535975d823ffef7e1686749b69b40347a8ec
2023-06-05 16:43:44 +08:00
hiyouga
a817801c0f tiny fix
Former-commit-id: 3c5da617cdab34c6cae038e3a06d0468ae4c6c86
2023-06-05 15:25:22 +08:00
hiyouga
063a83ab4e tiny fix
Former-commit-id: 5ce3e0056948aded120b63e365a892f9d8c3c840
2023-06-04 16:35:50 +08:00
hiyouga
eebe71699b tiny fix
Former-commit-id: a98ebf62fb82ffe5aaaea6a1ce3d4c60d23a5728
2023-06-04 12:55:40 +08:00
hiyouga
5f44112cf5 support QLoRA
Former-commit-id: d89597e28fe9b91246e58c55eeb9082436940481
2023-06-04 00:08:56 +08:00
hiyouga
2308d5a179 fix int8 inference
Former-commit-id: d05202943e9634526f96d189288f67852d3d1c40
2023-06-03 23:22:05 +08:00
hiyouga
7d6542115c reduce repetition penalty
Former-commit-id: 0381d93fc10ea5346724fe6295caa565a8eb4f61
2023-06-03 21:57:39 +08:00
hiyouga
c68f9ec3a9 fix int8 inference
Former-commit-id: fcf3506bef28504dd679c2210bdc84e5868e05fe
2023-06-03 21:17:47 +08:00
hiyouga
fa850ae6e5 add ziya prompt template
Former-commit-id: 321e44ac54a91260cf00a4caa1991708814473fc
2023-06-03 19:05:51 +08:00
hiyouga
5eef8d5d98 use low_cpu_mem_usage to speed up loading
Former-commit-id: 7891e4c200566a4a47088e93efd1fbebcb46528e
2023-06-03 18:19:01 +08:00
hiyouga
9b8b6623ac add logits processor
Former-commit-id: f6f4b1554ae1e8849b437d705ffa34ce7ebd56bb
2023-06-03 16:34:54 +08:00
hiyouga
ec48d06b9e remove unused code
Former-commit-id: f5c784dd75e152edb162ea02bf601d279f9de893
2023-06-03 00:10:54 +08:00
hiyouga
217b89cf7e add wechat
Former-commit-id: 0251b1788081a81773d87ca5ca2bac8428489852
2023-06-02 21:47:10 +08:00
hiyouga
382afc3822 tiny fix
Former-commit-id: d479aaf07a0997737da90c00558c42b32109928f
2023-06-02 19:02:25 +08:00
hiyouga
09997a25d3 fix layer norm name in PPO
Former-commit-id: 3ffc2efb997b717b3efad92a507584276e4bdfa1
2023-06-02 17:30:01 +08:00
hiyouga
58c8b29913 fix #1
Former-commit-id: b8e556e4632a79251a2225075c570337eeafa559
2023-06-02 14:25:00 +08:00
hiyouga
e9ab06678f alter rewards data type
Former-commit-id: 3eb7eb2d37525da50fe401ab7c59532e6e1ef984
2023-06-02 14:19:51 +08:00
hiyouga
896dbfec16 fix possibly OOM error
Former-commit-id: 0d590dffb41b0e832d9f87d20a23bcd0acd983aa
2023-06-01 23:54:44 +08:00
hiyouga
1512711ca2 fix bug at inference
Former-commit-id: df9b41af4401006b8040eb53c44dd290b604e0eb
2023-05-31 18:11:53 +08:00
hiyouga
a79df3500b update readme
Former-commit-id: 4054e85c664c541f435619baebcd687f80445d4a
2023-05-31 16:57:43 +08:00
hiyouga
693c049eac support BLOOM models
Former-commit-id: 1314b6ea39a01aa8ac325e1d875ac013d43aec45
2023-05-31 16:54:06 +08:00
hoshi-hiyouga
7492e8f208 Merge pull request #1 from mMrBun/main
Support conversation via API.

Former-commit-id: 5e64c7446718845444a14c5643c1bc5819562d60
2023-05-30 16:34:00 +08:00
hiyouga
181c776b58 remove dummy code
Former-commit-id: e6bc89d280945bbf48281107145c40a41d7cbd56
2023-05-30 16:28:00 +08:00
mMrBun
ef0aceaa50 Support conversation via API.
Former-commit-id: 57cfe9128ce1781853555b22b502ec85b8b01941
2023-05-30 15:00:28 +08:00
mMrBun
a18c6c0560 Support conversation via API.
Former-commit-id: 4d9d0ea083c15fb470ecbb428cb79b6dd48e3e92
2023-05-30 14:46:22 +08:00
hiyouga
b6ed5176e1 update readme
Former-commit-id: 64f7fa45a4dbd173a3e5eb66d17044aa1243cc4b
2023-05-29 21:54:01 +08:00
hiyouga
bda71e579b update readme
Former-commit-id: eed2773df8081b229d13b1d679bf2913715f23ac
2023-05-29 21:53:02 +08:00
hiyouga
33fee45217 add pre-training script
Former-commit-id: 935d58de2b3a2eadc4f0bed28c3ad7dee32e9fd5
2023-05-29 21:37:22 +08:00
hiyouga
304be6dc28 fix checkpoint loading
Former-commit-id: d31aa5c2c0bcb6a4ef4a62e21693548dd9acaae6
2023-05-29 17:43:16 +08:00
hiyouga
35d04a2c05 tiny fix
Former-commit-id: eae79707d31fd8be2cf4bee4d610557bbd49f6e7
2023-05-29 09:42:29 +08:00
hiyouga
83fc73c580 tiny fix
Former-commit-id: 08f7e0862b9df353a0e4d8274617c1a5e6fa6619
2023-05-28 21:48:33 +08:00
hiyouga
1fc551e1be use fp16 model, add logcallback
Former-commit-id: bea275d51338b49ce855eec0178e759607265e3d
2023-05-28 21:30:28 +08:00
hiyouga
17024ebc1a Initial commit
Former-commit-id: 5ca8e1d63727e7bcb8cab16542c763c47e48184a
2023-05-28 18:09:04 +08:00