Commit Graph

  • ba303fd1aa adapt for badam with ds zero3 Jonery 2024-06-17 18:18:10 +08:00
  • 60d9896a70 fix #4326 hiyouga 2024-06-17 18:17:48 +08:00
  • 485a80d294 tiny fix hiyouga 2024-06-17 17:47:25 +08:00
  • 63bfe9967e Update requirements.txt 胡翀 2024-06-17 16:45:57 +08:00
  • a720b82e63 Fix Dockerfile Eli Costa 2024-06-16 19:16:23 -03:00
  • d3b0048d8c Update README_zh.md Eli Costa 2024-06-16 11:34:31 -03:00
  • 9a0aca42a5 Update README_zh.md Eli Costa 2024-06-16 11:22:06 -03:00
  • 5e802b0645 Update README.md Eli Costa 2024-06-16 11:19:25 -03:00
  • dd7a1dbfae update packing with sdpa and eager attention mode ancv 2024-06-16 02:25:47 +07:00
  • ca67b7a568 Update parser.py hoshi-hiyouga 2024-06-16 02:57:00 +08:00
  • 76cd879c84 update pr template hiyouga 2024-06-16 01:43:43 +08:00
  • e0c049e590 Merge pull request #4307 from hiyouga/pissa hoshi-hiyouga 2024-06-16 01:41:50 +08:00
  • 727943f078 fix tol hiyouga 2024-06-16 01:38:44 +08:00
  • 8393b08666 Update tests.yml hiyouga 2024-06-16 01:22:23 +08:00
  • 9049f72d2f increase tol hiyouga 2024-06-16 01:21:06 +08:00
  • 32f45c9e91 support pissa hiyouga 2024-06-16 01:08:12 +08:00
  • 05f3a3c944 tiny fix hiyouga 2024-06-16 01:06:41 +08:00
  • f91fe10985 remove some unused params ancv 2024-06-15 23:00:55 +07:00
  • 14f7bfc545 use fixture hiyouga 2024-06-15 20:06:17 +08:00
  • 7f90b0cd20 add tests hiyouga 2024-06-15 19:51:20 +08:00
  • 308abfec6c add minicpm #4227 hiyouga 2024-06-15 17:58:52 +08:00
  • bb88536166 add license hiyouga 2024-06-15 17:54:33 +08:00
  • d2df3f2d6e update readme hiyouga 2024-06-15 05:13:16 +08:00
  • 2abfad9c1f fix #4271 hiyouga 2024-06-15 05:11:33 +08:00
  • 2af932d969 disable DP hiyouga 2024-06-15 04:57:19 +08:00
  • c29fa61a9c fix #4292 hiyouga 2024-06-15 04:47:13 +08:00
  • a30931fe0f fix #4295 hiyouga 2024-06-15 04:34:55 +08:00
  • 3ff9b87012 add test cases hiyouga 2024-06-15 04:05:54 +08:00
  • f4f315fd11 Update README.md hiyouga 2024-06-13 16:02:21 +08:00
  • 530165d9a5 update examples hiyouga 2024-06-13 03:26:10 +08:00
  • dbd1458adf add quant check in webui export tab hiyouga 2024-06-13 03:19:18 +08:00
  • dedefecd2b Update llama3_full_sft_ds3.yaml hiyouga 2024-06-13 03:16:20 +08:00
  • 46f441dd37 update examples hiyouga 2024-06-13 03:15:06 +08:00
  • 49b58fd6af fix #4221 hiyouga 2024-06-13 02:48:21 +08:00
  • 103a507b39 fix #4209 hiyouga 2024-06-13 02:25:50 +08:00
  • 0a75224f62 clean code hiyouga 2024-06-13 01:58:16 +08:00
  • 04d7629abf Merge pull request #4246 from hzhaoy/adapt-vllm-v0.5.0 hoshi-hiyouga 2024-06-13 01:54:02 +08:00
  • 1b6786a21f add neo-sft dataset hiyouga 2024-06-13 01:00:56 +08:00
  • 5080f2314c fix lint hiyouga 2024-06-13 00:48:44 +08:00
  • 41beb7f0a3 fix docker compose usage hiyouga 2024-06-13 00:07:48 +08:00
  • 799873aa14 adapt vllm==0.5.0 hzhaoy 2024-06-12 18:29:03 +08:00
  • fe2c7eaa93 update readme hiyouga 2024-06-12 17:39:12 +08:00
  • 6392d45ea7 fix #4242 hiyouga 2024-06-12 16:50:11 +08:00
  • c60ea675d7 Merge pull request #4234 from kimdwkimdw/patch-1 hoshi-hiyouga 2024-06-12 16:39:09 +08:00
  • 16c7c92396 Support vllm==0.5.0 Arthur Kim 2024-06-12 16:49:12 +09:00
  • c7ab302c69 implement efficient packing without cross-contamination attention ancv 2024-06-12 11:56:01 +07:00
  • 7598b37543 Merge pull request #4204 from dignfei/main hoshi-hiyouga 2024-06-11 17:06:10 +08:00
  • cc9717e2f2 Update pretrain.py hoshi-hiyouga 2024-06-11 17:02:14 +08:00
  • 08f2f99f4b fix deepspeed version hiyouga 2024-06-11 16:52:36 +08:00
  • 77bf3d66c7 经过大量的增量预训练,进行对比试验,发现这个bug:llama3在预训练时使用的tokenizer.eos_toke是'<|end_of_text|>' ,这里在每条数据后面也得用这个,而不是'<|eot_id|>',否则很容易导致严重的性能下降 d 2024-06-11 16:21:48 +08:00
  • f14f67f803 Update bug-report.yml hiyouga 2024-06-11 15:40:21 +08:00
  • 820b6e7b32 fix #4198 hiyouga 2024-06-11 15:38:38 +08:00
  • 27aece94cf tiny fix hiyouga 2024-06-11 12:48:53 +08:00
  • 3f2508be92 Merge pull request #4191 from iamthebot/al--add_manifest_for_reqs hoshi-hiyouga 2024-06-11 10:41:15 +08:00
  • fce11bb386 add manifest so requirements.txt in sdist Alfredo Luque 2024-06-11 00:07:06 +00:00
  • 2723438531 tiny fix hiyouga 2024-06-11 01:04:16 +08:00
  • f330b73682 set dev version hiyouga 2024-06-11 00:50:53 +08:00
  • 0f1e592326 release v0.8.1 v0.8.1 hiyouga 2024-06-11 00:44:26 +08:00
  • 4d7dd0330d fix #4160 hiyouga 2024-06-11 00:37:17 +08:00
  • ea2ca2777f fix #4145 hiyouga 2024-06-11 00:19:17 +08:00
  • 4b2b92fd9a update evaluator hiyouga 2024-06-10 23:56:00 +08:00
  • 784088db3f fix #2666 hiyouga 2024-06-10 21:24:15 +08:00
  • 0ecf0d51e3 Merge pull request #4167 from yzoaim/branch hoshi-hiyouga 2024-06-10 16:24:33 +08:00
  • bc04ca464a Optimize the handling of QWEN2 in scenarios involving multiple tool calls. mMrBun 2024-06-10 02:00:14 +08:00
  • 44829df762 Removed unnecessary comments. mMrBun 2024-06-09 18:25:22 +08:00
  • 94ddfa66c0 Merge branch 'hiyouga:main' into main mMrBun 2024-06-09 18:17:24 +08:00
  • 8db8ed5a41 Implemented the tool_formatter and tool_extractor for glm4 tool_format mMrBun 2024-06-09 18:16:15 +08:00
  • 041ecd0de1 fix README -.- 2024-06-08 23:51:56 +08:00
  • d812249db7 add pr ci hiyouga 2024-06-08 21:25:35 +08:00
  • 88528f1a87 Update tests.yml hiyouga 2024-06-08 21:15:36 +08:00
  • 82533114a7 update git workflows hiyouga 2024-06-08 21:11:32 +08:00
  • 6d9fbb3fa9 fix llamafactory-cli env hiyouga 2024-06-08 07:15:45 +08:00
  • 9953ae3d03 set dev version hiyouga 2024-06-08 06:46:09 +08:00
  • c0c387e4db release v0.8.0 v0.8.0 hiyouga 2024-06-08 05:20:54 +08:00
  • ae60ea15da add ultrafeedback and fineweb #4085 #4132 hiyouga 2024-06-08 02:42:34 +08:00
  • 72cd1123a8 fix ci hiyouga 2024-06-08 02:00:44 +08:00
  • 1364190a66 fix ci hiyouga 2024-06-08 01:57:36 +08:00
  • 6d17c59090 add ci hiyouga 2024-06-08 01:48:30 +08:00
  • e0f2c0b5dc init unittest hiyouga 2024-06-08 01:35:58 +08:00
  • 073e34855d Delete .readthedocs.yaml hiyouga 2024-06-08 00:58:10 +08:00
  • ff9ba70bb8 reorganize adapter code hiyouga 2024-06-08 00:47:23 +08:00
  • adbebb0e3f fix #4139 hoshi-hiyouga 2024-06-08 00:45:02 +08:00
  • 3f6b3eed98 add resume args in webui hiyouga 2024-06-08 00:22:16 +08:00
  • f45e81e186 fix #4137 hiyouga 2024-06-07 19:16:06 +08:00
  • ba648fd003 tiny fix hiyouga 2024-06-07 05:19:21 +08:00
  • b0e5a76f4c fix ppo trainer save zero3 model hiyouga 2024-06-07 05:14:19 +08:00
  • 8692796c9b fix ppo in trl 0.8.6 hiyouga 2024-06-07 04:48:29 +08:00
  • d0edcde4ea fix #4120 hiyouga 2024-06-07 04:18:05 +08:00
  • 8c4c2e580c update data processors hiyouga 2024-06-07 04:15:40 +08:00
  • 07f33e7641 Merge pull request #4009 from AlongWY/main hoshi-hiyouga 2024-06-07 03:48:46 +08:00
  • 1998c641af Update supervised.py hoshi-hiyouga 2024-06-07 03:42:08 +08:00
  • be1e5f9d62 Update supervised.py hoshi-hiyouga 2024-06-07 03:38:23 +08:00
  • fdeec6db52 Update supervised.py hoshi-hiyouga 2024-06-07 03:38:04 +08:00
  • a4d335b42f add qwen2 models hiyouga 2024-06-07 00:22:57 +08:00
  • fcb134e144 rename files hiyouga 2024-06-07 00:09:06 +08:00
  • a47e24222a add DISABLE_TORCHRUN option hiyouga 2024-06-06 23:44:58 +08:00
  • b96b995620 Merge pull request #4082 from MengqingCao/bugfix hoshi-hiyouga 2024-06-06 23:38:40 +08:00
  • c231706aa5 Update cli.py hoshi-hiyouga 2024-06-06 23:38:09 +08:00
  • 35b5117a59 fix ppo+zero3 #3108 hiyouga 2024-06-06 23:30:07 +08:00
  • 80f716bc10 fix torch gc hiyouga 2024-06-06 20:30:25 +08:00