Commit Graph

  • 6b6430489a fix loading best model hiyouga 2023-06-28 01:55:12 +08:00
  • 4ae8a20e1d fix RM accuracy hiyouga 2023-06-28 01:40:13 +08:00
  • eca15bf252 add star history hiyouga 2023-06-27 23:56:29 +08:00
  • e19dcc13e3 tiny fix hiyouga 2023-06-27 23:54:24 +08:00
  • 2d22961c7d fix initializing data arguments hiyouga 2023-06-27 22:50:23 +08:00
  • 640f774d30 support save full model, replace BOS token hiyouga 2023-06-27 21:40:11 +08:00
  • 33c2b063c6 fix decoding in seq2seq hiyouga 2023-06-27 19:33:08 +08:00
  • a7e53dcfef Update evaluate.py hiyouga 2023-06-26 23:41:33 +08:00
  • fe7ca5cb63 Create evaluate.py hiyouga 2023-06-26 23:30:18 +08:00
  • 0ff82b1304 Merge pull request #86 from Jingsong-Yan/main hoshi-hiyouga 2023-06-26 20:14:40 +08:00
  • d2de3f9e41 Update README.md with baichuan-7b-rtx3090 Jingsong-Yan 2023-06-26 19:45:41 +08:00
  • d5260ea860 Merge branch 'main' of https://github.com/hiyouga/LLaMA-Efficient-Tuning hiyouga 2023-06-26 18:07:09 +08:00
  • a8f580d753 fix generation in seq2seq.py hiyouga 2023-06-26 18:07:06 +08:00
  • 20c1b25ad9 Merge pull request #84 from wu-yy/patch-1 hoshi-hiyouga 2023-06-26 15:39:08 +08:00
  • 88840c4f2b Update requirements.txt 蓝鲸123 2023-06-26 15:36:19 +08:00
  • 3aa1ca66e0 support prefixes, loading multiple local files hiyouga 2023-06-26 15:32:40 +08:00
  • 83346e86af update api hiyouga 2023-06-26 13:39:57 +08:00
  • f9332bc329 update readme hiyouga 2023-06-23 00:17:05 +08:00
  • 7daf6c8b8e update API hiyouga 2023-06-22 20:46:24 +08:00
  • 391bf1c699 match api with OpenAI format hiyouga 2023-06-22 20:27:00 +08:00
  • 84b66010a3 Merge pull request #68 from mMrBun/main hoshi-hiyouga 2023-06-22 15:52:34 +08:00
  • 810d9e36ea Compatible with OpenAI API. Bun 2023-06-21 14:45:04 +08:00
  • de2c418637 add default template hiyouga 2023-06-16 21:12:17 +08:00
  • 7dc1f06a97 add belle multiturn dataset hiyouga 2023-06-16 20:01:16 +08:00
  • ee22b80ad0 fix freeze layers hiyouga 2023-06-16 17:38:21 +08:00
  • de9da40b18 add source prefix hiyouga 2023-06-16 16:32:17 +08:00
  • 3836aadacf support loading lora from hub hiyouga 2023-06-16 00:02:17 +08:00
  • 194c5d2bee support baichuan model hiyouga 2023-06-15 16:02:01 +08:00
  • 496846e819 fix bug in template vanilla hiyouga 2023-06-15 14:36:55 +08:00
  • c42562d7ae add BOS token in pre-training hiyouga 2023-06-15 01:46:17 +08:00
  • aa1bb8a9a2 support multiturn training like FastChat hiyouga 2023-06-14 22:27:39 +08:00
  • 6f655e3916 fix loading valuehead hiyouga 2023-06-13 11:13:06 +08:00
  • 6828f07d54 fix generating args hiyouga 2023-06-13 01:33:56 +08:00
  • 4724ae3492 support RM metrics, add generating Args hiyouga 2023-06-12 15:48:48 +08:00
  • 4c5cad9722 Merge pull request #26 from BUAADreamer/main hoshi-hiyouga 2023-06-11 19:06:29 +08:00
  • 4adbb95b03 add some BUAADreamer 2023-06-11 18:55:53 +08:00
  • 5b93ca6c39 add code for reading from multi files in one directory BUAADreamer 2023-06-10 16:27:30 +08:00
  • ef6c5ae18a add code for reading from multi files in one directory BUAADreamer 2023-06-10 15:53:47 +08:00
  • 03c92c79ff tiny fix hiyouga 2023-06-07 16:42:31 +08:00
  • fc6091e118 tiny fix hiyouga 2023-06-07 16:02:07 +08:00
  • 025670b4f6 tiny fix hiyouga 2023-06-07 12:58:14 +08:00
  • d6b32dd9ea add templates hiyouga 2023-06-07 12:40:44 +08:00
  • f57dae4a1a add belle template hiyouga 2023-06-07 12:30:11 +08:00
  • 5e2ec2d104 tiny fix hiyouga 2023-06-07 12:08:39 +08:00
  • b9feb82e4e add prompt template class hiyouga 2023-06-07 11:55:25 +08:00
  • 3da427a665 fix inference, add prompt template hiyouga 2023-06-07 10:52:35 +08:00
  • 12094c1db5 recover logging hiyouga 2023-06-06 21:36:37 +08:00
  • bf5ad34196 support distributed quantized training hiyouga 2023-06-06 17:39:41 +08:00
  • ac6f50dedf add API demo from #1 hiyouga 2023-06-05 21:32:18 +08:00
  • 8fd9ef924d Merge pull request #11 from hiyouga/api hoshi-hiyouga 2023-06-05 20:58:02 +08:00
  • a409e1f42c fix bug in web demo hiyouga 2023-06-05 17:58:29 +08:00
  • 3f5869111b increase max length in cli demo hiyouga 2023-06-05 16:49:14 +08:00
  • f9c51a8340 implement stream generating hiyouga 2023-06-05 16:43:44 +08:00
  • a817801c0f tiny fix hiyouga 2023-06-05 15:25:22 +08:00
  • 063a83ab4e tiny fix hiyouga 2023-06-04 16:35:50 +08:00
  • eebe71699b tiny fix hiyouga 2023-06-04 12:55:40 +08:00
  • 5f44112cf5 support QLoRA hiyouga 2023-06-04 00:08:56 +08:00
  • 2308d5a179 fix int8 inference hiyouga 2023-06-03 23:22:05 +08:00
  • 7d6542115c reduce repetition penalty hiyouga 2023-06-03 21:57:39 +08:00
  • c68f9ec3a9 fix int8 inference hiyouga 2023-06-03 21:17:47 +08:00
  • fa850ae6e5 add ziya prompt template hiyouga 2023-06-03 19:05:51 +08:00
  • 5eef8d5d98 use low_cpu_mem_usage to speed up loading hiyouga 2023-06-03 18:19:01 +08:00
  • 9b8b6623ac add logits processor hiyouga 2023-06-03 16:34:54 +08:00
  • ec48d06b9e remove unused code hiyouga 2023-06-03 00:10:54 +08:00
  • 217b89cf7e add wechat hiyouga 2023-06-02 21:47:10 +08:00
  • 382afc3822 tiny fix hiyouga 2023-06-02 19:02:25 +08:00
  • 09997a25d3 fix layer norm name in PPO hiyouga 2023-06-02 17:30:01 +08:00
  • 58c8b29913 fix #1 hiyouga 2023-06-02 14:25:00 +08:00
  • e9ab06678f alter rewards data type hiyouga 2023-06-02 14:19:51 +08:00
  • 896dbfec16 fix possibly OOM error hiyouga 2023-06-01 23:54:44 +08:00
  • 1512711ca2 fix bug at inference hiyouga 2023-05-31 18:11:53 +08:00
  • a79df3500b update readme hiyouga 2023-05-31 16:57:43 +08:00
  • 693c049eac support BLOOM models hiyouga 2023-05-31 16:54:06 +08:00
  • 7492e8f208 Merge pull request #1 from mMrBun/main hoshi-hiyouga 2023-05-30 16:34:00 +08:00
  • 181c776b58 remove dummy code hiyouga 2023-05-30 16:28:00 +08:00
  • ef0aceaa50 Support conversation via API. mMrBun 2023-05-30 15:00:28 +08:00
  • a18c6c0560 Support conversation via API. mMrBun 2023-05-30 14:46:22 +08:00
  • b6ed5176e1 update readme hiyouga 2023-05-29 21:54:01 +08:00
  • bda71e579b update readme hiyouga 2023-05-29 21:53:02 +08:00
  • 33fee45217 add pre-training script hiyouga 2023-05-29 21:37:22 +08:00
  • 304be6dc28 fix checkpoint loading hiyouga 2023-05-29 17:43:16 +08:00
  • 35d04a2c05 tiny fix hiyouga 2023-05-29 09:42:29 +08:00
  • 83fc73c580 tiny fix hiyouga 2023-05-28 21:48:33 +08:00
  • 1fc551e1be use fp16 model, add logcallback hiyouga 2023-05-28 21:30:28 +08:00
  • 17024ebc1a Initial commit hiyouga 2023-05-28 18:09:04 +08:00