wassname/peft
1
0
Fork 0
mirror of https://github.com/wassname/peft.git synced 2026-09-26 14:00:37 +08:00
Code Issues Packages Projects Releases Wiki Activity
245 Commits 1 Branch 0 Tags
48dc4c624e9540d24af41bcde14e66ce744a206b
Commit Graph
245 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
alvanli 48dc4c624e Add callback to save to local 2023-03-08 09:57:13 -05:00
Sourab Mangrulkar 8358b27445 Merge pull request #149 from huggingface/smangrul/fixes 
minor fixes to the examples
 2023-03-07 14:08:25 +05:30
Sourab Mangrulkar b9451ab458 fixing issues and quality ✨ 2023-03-07 14:04:19 +05:30
Sourab Mangrulkar ce4e6f3dd9 Merge pull request #150 from mayank31398/mayank/single-module 
support option for encoder only prompts
 2023-03-04 09:03:26 +05:30
mayank31398 53eb209387 support option for encoder only prompts 2023-03-03 23:43:25 +05:30
Sourab Mangrulkar a84414f6de minor fixes to the examples 2023-03-03 19:36:13 +05:30
Sourab Mangrulkar 2c532713ad Merge pull request #125 from SauravMaheshkar/minimal-structure 
chore: update `pyproject.toml`
 2023-03-02 19:07:59 +05:30
Saurav Maheshkar 94f00b7d27 chore: update Makefile with ruff commands 2023-02-28 10:46:07 +00:00
Saurav MaheshkarandSourab Mangrulkar 7820a539dd fix(pyproject.toml): update known_first_party 
Co-authored-by: Sourab Mangrulkar <13534540+pacman100@users.noreply.github.com>
 2023-02-28 16:10:38 +05:30
Saurav Maheshkar 47601bab7c chore: drop setup.cfg as we shifted to ruff 2023-02-28 02:58:12 +05:30
Saurav Maheshkar 99901896cc style: switch to ruff 2023-02-27 10:50:10 +00:00
Saurav Maheshkar 5c7fe97753 Merge branch 'huggingface:main' into minimal-structure 2023-02-27 10:37:41 +00:00
Sourab Mangrulkar aa18556c56 Merge pull request #140 from zanussbaum/zero_init 
fix: count params when zero init'd
 2023-02-27 13:18:30 +05:30
zanussbaum e6bf09db80 fix: count params when zero init'd 2023-02-26 22:31:20 -05:00
Sourab Mangrulkar 681ce93cc1 Merge pull request #134 from gabinguo/main 
issue#126: torch.load device issue.
 2023-02-25 11:43:43 +05:30
Kunpeng GUO 85ad682530 issue#126: torch.load device issue. 2023-02-25 07:09:07 +01:00
Sourab Mangrulkar e19ee681ac Merge pull request #127 from huggingface/smangrul/make-activation-checkpointing-optional 
fixes `prepare_for_int8_training`
 2023-02-25 02:13:04 +05:30
Sourab Mangrulkar 83d6d55d4b address layernorm issue 2023-02-24 18:24:27 +05:30
Sourab Mangrulkar 7dfb472424 make gradient checkpointing optional when using PEFT+INT8 2023-02-24 13:02:40 +05:30
Saurav Maheshkar a78f8a0495 style: move isort and pytest config to pyproject 2023-02-23 11:34:18 +00:00
Saurav Maheshkar 6175ee2c4c chore: drop MANIFEST 2023-02-23 11:30:14 +00:00
Younes Belkada a3537160dc fix autocast issue (#121) 2023-02-23 09:17:40 +01:00
Sourab Mangrulkar 75925b1aae Merge pull request #117 from huggingface/smangrul/fix-lora-merging-in-inference 
fix merging lora weights for inference
 2023-02-22 00:20:40 +05:30
Sourab Mangrulkar 1ef0f89a0c add util for getting the base model 2023-02-22 00:14:24 +05:30
Sourab Mangrulkar e6ef85a711 fix merging lora weights for inference 2023-02-22 00:00:36 +05:30
Sourab Mangrulkar 6f2803e8a7 Merge pull request #109 from huggingface/smangrul/add-gpt-neox 
add `EleutherAI/gpt-neox-20b` to support matrix
 2023-02-18 12:38:04 +05:30
Sourab Mangrulkar 1c9d197693 add EleutherAI/gpt-neox-20b to support matrix 2023-02-18 12:37:02 +05:30
Sourab Mangrulkar 592b1dd99f Merge pull request #106 from huggingface/smangrul/add-diable-adapter-context-manager 
add disable adapter context manager
 2023-02-17 20:32:31 +05:30
Sourab Mangrulkar 3240c0bb36 Merge pull request #107 from mrm8488/patch-1 
Fix typo
 2023-02-17 20:30:27 +05:30
Manuel Romero e8fbcfcac3 Fix typo 2023-02-17 15:08:47 +01:00
Sourab Mangrulkar 1a8928c5a4 Update lora.py 2023-02-17 17:48:16 +05:30
Sourab Mangrulkar 173dc3dedf add disable_adapter context manager 2023-02-17 17:40:45 +05:30
Younes Belkada dbf44fe316 [core] Some changes with prepare_model_for_training & few fixes (#105) 
* changes

* apply to other notebooks
 2023-02-17 10:49:11 +01:00
Sourab Mangrulkar 648fcb397c Merge pull request #104 from huggingface/smangrul/make_lora_target_modules_accept_regex 
add support for regex target modules in lora
 2023-02-17 15:00:58 +05:30
Sourab Mangrulkar 7aadb6d9ec add support for regex target modules in lora 2023-02-17 14:52:03 +05:30
Sourab Mangrulkar 49842e1961 Merge pull request #97 from huggingface/smangrul/make-bnb-optional 
making `bnb` optional
 2023-02-16 22:33:51 +05:30
Sourab Mangrulkar 44d0ac3f25 fix 2023-02-16 20:15:48 +05:30
Sourab Mangrulkar 43a9a42991 fix 2023-02-16 20:11:39 +05:30
Sourab Mangrulkar 145b13c238 making bnb optional 2023-02-16 20:07:06 +05:30
Sourab Mangrulkar 8ace5532b2 Merge pull request #95 from huggingface/smangrul/add-whisper-example 
adding whisper large peft+int8 training example
 2023-02-16 17:43:46 +05:30
Sourab Mangrulkar c1281b96ff resolving comments and running jupyter black 2023-02-16 17:42:28 +05:30
Sourab Mangrulkar ca7b46209a adding whisper large peft+int8 training example 2023-02-16 15:06:42 +05:30
Sourab Mangrulkar 81285f30a5 Merge pull request #90 from huggingface/smangrul/fix-prepare-inputs-for-training 
making `prepare_model_for_training` flexible
 2023-02-16 11:38:35 +05:30
Sourab Mangrulkar c9b225d257 revert 2023-02-15 16:58:33 +05:30
Sourab Mangrulkar af7414a67d fix forward signature 2023-02-15 15:44:26 +05:30
Sourab Mangrulkar 6d6149cf81 preventing other 1D layers to be casted in FP32 2023-02-15 14:03:47 +05:30
Younes Belkada a31dfa3001 Merge pull request #86 from younesbelkada/add-flan-t5-int8 
[`bnb`] add flan-t5 example
 2023-02-15 09:02:24 +01:00
younesbelkada afa7739131 update 2023-02-15 08:01:56 +00:00
Sourab Mangrulkar f1ee1e4c0f making prepare_model_for_training flexible 2023-02-15 12:51:23 +05:30
Sourab Mangrulkar ed5a7bff6b Merge pull request #85 from younesbelkada/int8-wrapper 
[`core`]  add `prepare_model_for_training`
 2023-02-15 11:21:03 +05:30