Commit Graph

  • 6eed1db36c Release v0.1.7 v0.1.7 hiyouga 2023-08-18 17:21:27 +08:00
  • 948124f55e tiny fix hiyouga 2023-08-18 13:07:35 +08:00
  • 2b191ca776 support ppo score norm (trl 0.5.1.dev required) hiyouga 2023-08-18 12:02:42 +08:00
  • be4d2822ea fix PPO trainer #551 , update readme hiyouga 2023-08-18 11:43:10 +08:00
  • 736ddd0319 update readme hiyouga 2023-08-18 01:51:55 +08:00
  • dfa289aa72 Update .gitignore hiyouga 2023-08-18 01:43:42 +08:00
  • c2644f939a update training resuming hiyouga 2023-08-18 01:41:17 +08:00
  • f11c1ae562 Merge pull request #434 from niuba/main hoshi-hiyouga 2023-08-18 01:38:31 +08:00
  • 3126164aa6 Merge branch 'main' into main hoshi-hiyouga 2023-08-18 01:37:23 +08:00
  • ed10486cad support bf16 ppo #551 hiyouga 2023-08-18 00:40:32 +08:00
  • 04fa430c6c fix ChatGLM2 ppo #527 #528 hiyouga 2023-08-18 00:34:59 +08:00
  • fa1893b59c fix generation bug #532 hiyouga 2023-08-17 22:21:34 +08:00
  • e993e717a5 fix streaming in pt stage #548 #549 hiyouga 2023-08-17 17:59:26 +08:00
  • c80e56423a update readme hiyouga 2023-08-17 11:00:22 +08:00
  • ffa09a01d6 fix baichuan and intern template hiyouga 2023-08-17 01:27:20 +08:00
  • 7d04f8567b fix generation hiyouga 2023-08-16 22:39:54 +08:00
  • baa709674f fix system prompt hiyouga 2023-08-16 01:35:52 +08:00
  • ca9a494d0c fix baichuan template #481 hiyouga 2023-08-15 11:38:21 +08:00
  • 37eb8c05cc Merge pull request #516 from liuyanyi/add_gitignore hoshi-hiyouga 2023-08-15 11:25:40 +08:00
  • 7c046edb7b fix ChatGLM RLHF hiyouga 2023-08-15 11:19:20 +08:00
  • 22cea38b20 Add .gitignore Yanyi Liu 2023-08-15 11:13:45 +08:00
  • ef2ca0a827 alert pad_token source hiyouga 2023-08-15 00:07:56 +08:00
  • 7f0b908de2 update webui hiyouga 2023-08-14 22:45:26 +08:00
  • 5fc5e776ff Merge pull request #511 from hiyouga/feature-autoTemplate hoshi-hiyouga 2023-08-14 22:44:04 +08:00
  • 93b281c016 auto match template when change model_name codemayq 2023-08-14 20:56:05 +08:00
  • 9585699918 add template match and stage in webui codemayq 2023-08-14 20:42:59 +08:00
  • bceaba551d fix ChatGLM lm_head #494 hiyouga 2023-08-14 14:14:48 +08:00
  • 0bfeed3a7e fix bug in webui hiyouga 2023-08-14 11:38:42 +08:00
  • 70a780c3c0 fix webui cache hiyouga 2023-08-14 11:37:01 +08:00
  • d74ab5306c update readme_zh hiyouga 2023-08-14 11:13:25 +08:00
  • 688e8601ab web UI integrating RLHF hiyouga 2023-08-14 10:48:47 +08:00
  • 4933ab5956 fix #480 hiyouga 2023-08-14 00:23:56 +08:00
  • 6c7225a5d4 fix webui hiyouga 2023-08-12 23:52:07 +08:00
  • a22982f2fa tiny fix hiyouga 2023-08-12 22:02:43 +08:00
  • c95479dddb fix rope scaling hiyouga 2023-08-12 22:00:01 +08:00
  • fc48bd8da0 update readme hiyouga 2023-08-12 21:29:06 +08:00
  • d5323bfa3f update readme hiyouga 2023-08-12 21:25:19 +08:00
  • e9d4a2b507 update readme hiyouga 2023-08-12 21:23:05 +08:00
  • 37bcbe8046 update readme hiyouga 2023-08-12 21:00:11 +08:00
  • fdfb644f0a support rope scaling, fix #475 #476 #478 hiyouga 2023-08-12 20:46:27 +08:00
  • cde9f3db57 Merge pull request #479 from hiyouga/feature-addCmdExport hoshi-hiyouga 2023-08-12 20:41:52 +08:00
  • 8bf5a98815 add sft script preview in webui codemayq 2023-08-12 13:53:55 +08:00
  • be566a15a5 fix unusual output of 8bit models #278 #391 hiyouga 2023-08-12 00:25:29 +08:00
  • d5f1b99ac4 Release v0.1.6 v0.1.6 hiyouga 2023-08-11 23:25:57 +08:00
  • 2144bb0e27 Update README_zh.md hiyouga 2023-08-11 14:06:02 +08:00
  • bc665bacc7 add defaults hiyouga 2023-08-11 13:56:26 +08:00
  • 52bfcf4883 fix stop word in baichuan template hiyouga 2023-08-11 13:51:46 +08:00
  • 06df3d6fb6 fix baichuan template hiyouga 2023-08-11 13:45:47 +08:00
  • ca719a8697 support DPO training (2305.18290) hiyouga 2023-08-11 03:02:53 +08:00
  • 72dfd74005 Merge pull request #451 from jovialchen/main hoshi-hiyouga 2023-08-10 17:25:38 +08:00
  • 69302c4420 fix webui val size hiyouga 2023-08-10 15:20:44 +08:00
  • 42d7019b2e huggingface login for projects must login while running jiongxuc 2023-08-10 14:57:12 +08:00
  • 5f0d0d6b9b fix template hiyouga 2023-08-09 23:14:27 +08:00
  • 76cb63e4f6 fix template hiyouga 2023-08-09 23:10:20 +08:00
  • 467d571206 support val set in streaming mode hiyouga 2023-08-09 23:00:26 +08:00
  • 972bfa700a fix tokenizer hiyouga 2023-08-09 17:52:15 +08:00
  • 458955d0fb add last_checkpoint support niuba 2023-08-09 16:39:27 +08:00
  • 990eeccf45 fix sft trainer hiyouga 2023-08-09 16:35:03 +08:00
  • a3a7465f00 fix rm #420, fix template #426, fix #423 hiyouga 2023-08-09 16:23:31 +08:00
  • 031a819257 fix llama2 template hoshi-hiyouga 2023-08-09 00:58:27 +08:00
  • eb4b4e3c8c fix tokenizer hoshi-hiyouga 2023-08-09 00:54:54 +08:00
  • d2e1fe9b1d update webui hiyouga 2023-08-09 00:26:11 +08:00
  • 6e27a9e39a fix tokenizer #417 hiyouga 2023-08-08 23:59:41 +08:00
  • 805478c911 fix bug hiyouga 2023-08-08 21:28:28 +08:00
  • a281cdeb89 fix bug hiyouga 2023-08-08 17:55:55 +08:00
  • cda698a67f fix chatml template #408 hiyouga 2023-08-08 17:44:39 +08:00
  • 15acd17716 update args spec hiyouga 2023-08-07 15:23:35 +08:00
  • 34a2bddfcd update readme hiyouga 2023-08-07 15:02:02 +08:00
  • 370f817549 Merge branch 'main' of https://github.com/hiyouga/LLaMA-Efficient-Tuning hiyouga 2023-08-07 13:59:16 +08:00
  • 041390c37e fix #376 hiyouga 2023-08-07 13:58:59 +08:00
  • d9fe4bf500 Merge pull request #382 from hiyouga/feature-updateReadme hoshi-hiyouga 2023-08-07 13:43:38 +08:00
  • e0c7e944fc update trainer hiyouga 2023-08-07 13:34:35 +08:00
  • 0845fe67db add detailed model configs codemayq 2023-08-07 09:30:23 +08:00
  • fe3b12d900 fix qwen eos token hiyouga 2023-08-06 13:31:17 +08:00
  • a70d56864e fix qwen tokenizer #361 hiyouga 2023-08-05 17:06:05 +08:00
  • fdbb2c5378 fix template for tiktoken hiyouga 2023-08-05 13:42:42 +08:00
  • 3c0aaf42af remove redundant code hiyouga 2023-08-05 00:27:27 +08:00
  • 438e19160a fix template hiyouga 2023-08-05 00:25:00 +08:00
  • f2b2ff6950 fix llama2 template hiyouga 2023-08-05 00:07:54 +08:00
  • 86cef96305 Support safe ChatML template, fix qwen tok #351 #354 hoshi-hiyouga 2023-08-05 00:00:23 +08:00
  • 5f50944baf fix bos and eos token hiyouga 2023-08-04 23:55:57 +08:00
  • 0804fd2353 fix encode hiyouga 2023-08-04 23:27:55 +08:00
  • 86419eb457 support chatml safe encoding hiyouga 2023-08-04 23:14:28 +08:00
  • 76f3ae7bf3 support interleave probs hiyouga 2023-08-04 21:27:35 +08:00
  • aaa85190eb fix webui export model hiyouga 2023-08-04 14:20:27 +08:00
  • e2a4e926b9 fix mtloader hiyouga 2023-08-03 19:29:02 +08:00
  • d6e922dc1c tiny fix hiyouga 2023-08-03 17:42:28 +08:00
  • 27f4317ec6 fix qwen inference hiyouga 2023-08-03 16:31:55 +08:00
  • e434348216 fix qwen inference hiyouga 2023-08-03 16:15:38 +08:00
  • 2e19afedb8 support Qwen-7B, fix InternLM-7B inference hiyouga 2023-08-03 15:53:32 +08:00
  • da08fa7c63 update web demo hiyouga 2023-08-03 13:28:28 +08:00
  • 9c96b97dc7 fix webui hiyouga 2023-08-03 12:43:12 +08:00
  • 28a51b622b modify code structure hiyouga 2023-08-02 23:17:36 +08:00
  • 8bd1da7144 fix PPO trainer hiyouga 2023-08-02 19:10:23 +08:00
  • e4d0b8ee6e update ppo trainer hiyouga 2023-08-02 18:46:41 +08:00
  • 1dfb28b362 fix memory leak of PPO trainer hiyouga 2023-08-02 17:41:34 +08:00
  • ba618947e7 release v0.1.5 v0.1.5 hiyouga 2023-08-02 16:10:31 +08:00
  • f81041b502 Merge pull request #307 from GitYCC/feature/fix-llama2-prompt-template hoshi-hiyouga 2023-08-02 15:51:28 +08:00
  • f2533a2800 [fix] Remove useless code YC Chen 2023-08-02 14:35:35 +08:00
  • bb5b4a7f26 [feature] Fix template of Llama2 to match the offical template YC Chen 2023-08-02 14:05:43 +08:00