tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Commit Graph

  • 34d3c8dcfb [docs] Update README Quickstart to use CLI (#2137) Wing Lian 2024-12-09 14:03:19 -05:00
  • ab4b32187d need to update deepspeed version in extras too (#2161) [skip ci] Wing Lian 2024-12-09 14:01:44 -05:00
  • 5d6b088997 fix: chat_template masking due to truncation, consolidate turn build and keys within field (#2123) [skip ci] NanoCode012 2024-12-10 01:49:38 +07:00
  • 3862267040 don't add dataset tags if empty due to all local data paths (#2162) [skip ci] Wing Lian 2024-12-09 13:49:18 -05:00
  • c78de6f214 feat: add kto example (#2158) [skip ci] NanoCode012 2024-12-09 20:17:27 +07:00
  • b1e8286c57 add missing __init__ to optimizers path (#2160) [skip ci] Wing Lian 2024-12-09 08:17:08 -05:00
  • 57a8c54682 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-09 12:27:34 +00:00
  • 40907c6887 upgrade deepspeed to 0.16.1 (#2157) Wing Lian 2024-12-09 07:25:10 -05:00
  • 6a342feda2 fix: duplicate mlflow logging (#2109) [skip ci] NanoCode012 2024-12-09 19:24:48 +07:00
  • 2de866e92f revert seq len to 8192 Sunny 2024-12-08 22:30:20 -05:00
  • 295e07dcca settings Sunny 2024-12-08 22:22:18 -05:00
  • 95d8c5aa3a Built site for gh-pages Quarto GHA Workflow Runner 2024-12-09 02:10:08 +00:00
  • 0c25bc07a2 use manual version for now (#2156) Wing Lian 2024-12-08 21:09:12 -05:00
  • 7e347cec5e Built site for gh-pages Quarto GHA Workflow Runner 2024-12-08 21:40:12 +00:00
  • 343a4d8855 Fixing issue#2134 Axolotl Crashes At The End Of Training If Base Model Is Local (#2140) Sunny Liu 2024-12-08 16:39:05 -05:00
  • 393853751e add additional fft deepspeed variants (#2153) [skip ci] Wing Lian 2024-12-08 16:38:47 -05:00
  • 39ab9626f1 add transformers module to cleanup e2e-fsdp-trainer Wing Lian 2024-12-08 14:52:54 -05:00
  • 26bd81cec0 re-enable tests w change in patching Wing Lian 2024-12-08 14:52:09 -05:00
  • d31729dc4e Built site for gh-pages Quarto GHA Workflow Runner 2024-12-08 19:51:36 +00:00
  • 1302e31049 Transformers version flexibility and FSDP optimizer patch (#2155) Wing Lian 2024-12-08 14:50:40 -05:00
  • 5dba7f2868 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-08 03:25:11 +00:00
  • be5f554a62 bump autoawq to 0.2.7.post3 (#2150) Wing Lian 2024-12-07 22:24:09 -05:00
  • 22319182ab fix for auto_map check when using remote code and multipack for models like deepseek (#2151) [skip ci] Wing Lian 2024-12-07 22:23:52 -05:00
  • 440aab8a6f add --version support to axolotl cli (#2152) [skip ci] Wing Lian 2024-12-07 22:23:33 -05:00
  • 51d722db83 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-07 22:25:38 +00:00
  • 3afc91fba9 run 2.5.1 test without waiting for 1st e2e docker-base-nvcr-pytorch Wing Lian 2024-12-07 08:47:36 -05:00
  • 0689419d25 use pr base tag Wing Lian 2024-12-07 05:05:02 -05:00
  • e64c32c0bd push test build Wing Lian 2024-12-06 20:15:02 -05:00
  • ec819dde3b attempt to build the test images Wing Lian 2024-12-06 19:50:08 -05:00
  • fdf4bb5087 fix default base image Wing Lian 2024-12-06 11:35:40 -05:00
  • f67d16268c try with default tag Wing Lian 2024-12-06 11:10:49 -05:00
  • 684b543aa1 experiment with nvcr pytorch image for torch 2.5.1 Wing Lian 2024-12-06 11:07:27 -05:00
  • 5bef19064b [tests] reset known modules that are patched on each test function end (#2147) Wing Lian 2024-12-07 17:24:46 -05:00
  • da3d8f31c6 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-07 10:03:54 +00:00
  • 743ba62bd5 Transformers 4.47.0 (#2138) Wing Lian 2024-12-07 05:03:01 -05:00
  • f9a7748bd8 Fix llama type model check (#2142) [skip ci] Chirag Jain 2024-12-07 15:32:32 +05:30
  • 31723ac523 fix whitespace for patch check transformers-4_47_0_v2 Wing Lian 2024-12-06 16:43:44 -05:00
  • 2e9e423dfd detab the code to check Wing Lian 2024-12-06 16:42:29 -05:00
  • cbe61186dc patches for llama ga Wing Lian 2024-12-06 16:40:24 -05:00
  • 3c07b6d6b1 lint bursteratom 2024-12-06 16:06:57 -05:00
  • 89dae7dc6d lora_target_module bursteratom 2024-12-06 15:41:09 -05:00
  • 1b54af8e54 lora config bursteratom 2024-12-06 15:27:18 -05:00
  • ca7b56cba3 lora config bursteratom 2024-12-06 15:26:06 -05:00
  • 2a83580bdc also bump accelerate Wing Lian 2024-12-06 15:24:57 -05:00
  • ea8269d2eb lora config bursteratom 2024-12-06 15:23:24 -05:00
  • 13ca7ed087 comment out lora target bursteratom 2024-12-06 15:21:08 -05:00
  • 0dfd8541ee lora config qwen2vl bursteratom 2024-12-06 14:56:51 -05:00
  • 75e1d3537f qwen2_vl get_text_config bursteratom 2024-12-06 14:54:06 -05:00
  • 825f66b9fd update HF HUB env var and fix reward trainer log since it doesn't directly override log Wing Lian 2024-12-06 14:52:59 -05:00
  • 2b7f3bd6ab qwen2_vl get_text_config bursteratom 2024-12-06 14:52:17 -05:00
  • d85a229afe get_text_config bursteratom 2024-12-06 14:50:05 -05:00
  • 355cd7c872 update is_multimodal requirement to include qwen2_vl bursteratom 2024-12-06 14:43:50 -05:00
  • eab1638686 lint bursteratom 2024-12-06 14:37:32 -05:00
  • a3a4d22709 config init qwen2-vl chat template bursteratom 2024-12-06 14:24:03 -05:00
  • f9eb7d8663 qwen2 example bursteratom 2024-12-06 14:22:08 -05:00
  • 343771a6d3 lint bursteratom 2024-12-06 13:15:49 -05:00
  • d2c32d0cba lint bursteratom 2024-12-06 13:04:42 -05:00
  • cec9887609 add llava chat template to config bursteratom 2024-12-06 12:57:20 -05:00
  • 88b2cae748 llava template bursteratom 2024-12-06 12:54:43 -05:00
  • 3b44989205 skip parent, call grandparent - yeah, super janky Wing Lian 2024-12-06 12:19:14 -05:00
  • aea2565938 for test only bursteratom 2024-12-06 11:54:07 -05:00
  • 811224d7b7 broken 🦥 with latest transformers Wing Lian 2024-12-06 11:34:06 -05:00
  • 84a14fc604 fix trl trainer.log interfaces Wing Lian 2024-12-06 10:35:29 -05:00
  • 86cf62ca46 fix: update trainer.log signature NanoCode012 2024-11-25 18:31:43 +07:00
  • fc54e10455 bump transformers and trl Wing Lian 2024-12-06 10:27:12 -05:00
  • 61dd11c23b Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 15:21:31 +00:00
  • cd7ff99173 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 15:20:46 +00:00
  • 5e9fa33f3d reduce test concurrency to avoid HF rate limiting, test suite parity (#2128) Wing Lian 2024-12-06 10:20:20 -05:00
  • 08fa133177 Fix broken CLI; remove duplicate metadata from setup.py (#2136) Dan Saunders 2024-12-06 10:19:54 -05:00
  • 9e7f6983de Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 14:09:51 +00:00
  • 6b3058b2dc upgrade bnb 0.45.0 and peft 0.14.0 (#2126) Wing Lian 2024-12-06 09:08:55 -05:00
  • 197389a388 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 03:38:24 +00:00
  • 5726141c4e remove accidentally included symlink (#2131) Wing Lian 2024-12-05 22:37:19 -05:00
  • 157ee933d1 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 03:13:49 +00:00
  • 0c807e1cb8 Built site for gh-pages Quarto GHA Workflow Runner 2024-12-06 03:12:45 +00:00
  • 2f3ebbc44f auto-versioning and adding axolotl.__version__ (#2127) Dan Saunders 2024-12-05 22:12:40 -05:00
  • fc973f4322 CLI Implementation with Click (#2107) Dan Saunders 2024-12-05 22:11:48 -05:00
  • e399ba533e fix license header for fix_untrained_tokens from unsloth-zoo (#2129) [skip ci] Wing Lian 2024-12-05 21:20:40 -05:00
  • 4baf8e5e96 cleanup the readme, add Modal as sponsor (#2130) [skip ci] Wing Lian 2024-12-05 21:19:52 -05:00
  • 1ad56303b2 lint bursteratom 2024-12-05 15:34:04 -05:00
  • dc055a4ef7 lint bursteratom 2024-12-05 14:59:51 -05:00
  • 169116a50f llava example bursteratom 2024-12-05 12:58:30 -05:00
  • 43e412f660 comment bursteratom 2024-12-04 13:18:25 -05:00
  • 7aa57803e1 fix optimizer reset for relora sft (#1414) Wing Lian 2024-12-03 08:58:23 -05:00
  • 1969fa3bf0 fix(readme): update cuda instructions during preprocess (#2114) [skip ci] NanoCode012 2024-12-03 20:58:03 +07:00
  • 4078f37076 feat: add cut_cross_entropy (#2091) NanoCode012 2024-12-03 20:22:22 +07:00
  • f073af6d99 fix merge conflict of duplicate max_steps in config for relora (#2116) Wing Lian 2024-12-03 07:42:41 -05:00
  • 139d2612fa fix so inference can be run against quantized models without adapters (#1834) Wing Lian 2024-12-03 00:02:38 -05:00
  • 20573fd13e Add ds model card, rebased (#2101) [skip ci] Sunny Liu 2024-12-03 00:02:02 -05:00
  • 2b7b4af81c fix(vlm): handle legacy conversation data format and check image in data (#2018) [skip ci] NanoCode012 2024-12-03 12:01:31 +07:00
  • d56260c8d5 Check torch version for ADOPT optimizer + integrating new ADOPT updates (#2104) Sunny Liu 2024-12-02 20:15:39 -05:00
  • cac785ec0e use pytest sugar and verbose for more info during ci (#2112) [skip ci] Wing Lian 2024-12-02 20:14:40 -05:00
  • e62991edef make the eval size smaller for the resume test (#2111) [skip ci] Wing Lian 2024-12-02 18:32:29 -05:00
  • fd9e7b55f6 build causal_conv1d and mamba-ssm into the base image (#2113) Wing Lian 2024-12-02 18:27:46 -05:00
  • c0c53eb62f various tests fixes for flakey tests (#2110) Wing Lian 2024-12-02 17:28:58 -05:00
  • b0fbd4d11d Add Exact Deduplication Feature to Preprocessing Pipeline (#2072) Oliver Molenschot 2024-12-02 05:47:10 -08:00
  • 1a70d4d6a4 add e2e tests for Unsloth qlora and test the builds (#2093) Wing Lian 2024-11-29 20:38:49 -05:00
  • d8787a433f support seperate lr for embeddings, similar to loraplus (#1910) [skip ci] Wing Lian 2024-11-29 20:38:20 -05:00
  • e775422269 fix: ds3 and fsdp lmbench eval (#2102) [ski[p ci] NanoCode012 2024-11-30 08:37:49 +07:00
  • 97178f5960 add finetome dataset to fixtures, check eval_loss in test (#2106) [skip ci] Wing Lian 2024-11-29 20:37:32 -05:00