ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
488 commits 391 branches 58 tags 304 MiB
Commit graph

152 commits

Author SHA1 Message Date
Roland Tannous
fbc934c231 Merge branch 'nightly' into feature/transformers-v5-support 2026-02-23 13:32:51 +00:00
Roland Tannous
d94f842158 fix: error on >30% sample drop after train_on_responses_only instead of silent DataLoader crash 2026-02-23 12:21:06 +00:00
Roland Tannous
2fbde1cf70 added shutil import to main.py 2026-02-23 07:44:59 +00:00
Roland Tannous
036d85c9e4 Merge nightly into feature/transformers-v5-support 2026-02-23 07:40:28 +00:00
Roland Tannous
198433363a feat: clear unsloth_compiled_cache on startup, shutdown, and between model loads 2026-02-23 07:26:22 +00:00
Roland Tannous
62c260a109 Add GLM, Qwen3 MoE, TinyQwen3 MoE, and Ministral 3 VL model defaults and GLM train_on_responses_only mapping 2026-02-23 05:51:43 +00:00
Roland Tannous
c281f2a3c6 Remove stale .venv_overlay on server startup to prevent transformers version conflicts 2026-02-23 05:08:27 +00:00
Roland Tannous
778762eb28 Patch adapter_config.json with unsloth_training_method and auto-detect load_in_4bit for LoRA inference 2026-02-22 20:27:52 +00:00
Roland Tannous
4d0f7c525b Purge own utils/core modules and use lazy imports so is_vision_model picks up fresh AutoConfig after version switch 2026-02-22 20:04:35 +00:00
Roland Tannous
cf245adb63 Add transformers version switch to model config and vision check endpoints for dropdown selection 2026-02-22 19:56:12 +00:00
Roland Tannous
332e071b6c Install transformers into both site-packages and overlay to fix sub-package resolution during version switch 2026-02-22 19:44:12 +00:00
Roland Tannous
7d2a8be0c0 Move transformers overlay to local .venv_overlay/, add huggingface-hub to overlay install 2026-02-22 19:34:46 +00:00
Roland Tannous
4d9ab493f7 Use sys.path overlay to switch transformers versions in-process instead of modifying site-packages 2026-02-22 19:19:12 +00:00
Roland Tannous
66aaaa9af7 Fix in-memory transformers version detection and aggressive module purge for 5.1.0/4.57.1 switching 2026-02-22 19:08:18 +00:00
Roland Tannous
2f212f95a2 aggressive reload_transformers 2026-02-22 18:52:06 +00:00
Roland Tannous
f5b30448e8 Auto-switch transformers version (5.1.0/4.57.1) for Ministral-3, GLM-4.7-Flash, Qwen3-30B-A3B models with LoRA adapter resolution 2026-02-22 18:29:40 +00:00
Roland Tannous
e866159e3b fix: correct vision LoRA defaults for VLMs and remove vision fields from text-only model configs 2026-02-22 15:09:10 +00:00
Roland Tannous
e666442b6e fix: pass full Processor as processing_class for VLM SFTTrainer 2026-02-22 14:11:12 +00:00
Roland Tannous
ef118d0d05 fix: load proper vision processor from base model when FastVisionModel returns raw tokenizer, add tokenize=False to vision chat template 2026-02-21 04:40:29 +00:00
Manan17
e9710874e1 Mapping proper tokenizer for VLMs 2026-02-21 01:57:05 +00:00
Manan17
756aa56cd2 fixed the vlm's text only errors 2026-02-20 22:23:26 +00:00
Roland Tannous
0d0038b901 Merge pull request #186 from unslothai/fix/colab-setup-fixes 
Fix/colab setup fixes
 2026-02-20 23:08:25 +04:00
Roland Tannous
e1a24c1ee9 add huggingface-hub==0.36.0 due to colab error 2026-02-20 18:26:20 +00:00
Roland Tannous
7c67228c40 add huggingface-hub==0.36.0 due to colab error 2026-02-20 18:24:48 +00:00
Roland Tannous
3bd0b8fb8b moved transformers4.57.1 to no-extra-deps 2026-02-20 18:08:58 +00:00
Manan17
bd0cee8c15 Setting it to total cpu_count // 4 2026-02-20 06:32:01 +00:00
Manan17
444ece6b07 Fixing compare feature 2026-02-19 20:15:44 +00:00
Roland Tannous
11b3029dc6 Simplify dataset check to 2-tier, improve multimodal detection, auto-set trainOnCompletions, recheck dataset on reload 2026-02-19 11:25:54 +00:00
Roland Tannous
58753151ef change train_on_completions to true 2026-02-19 06:02:16 +00:00
Manan17
2755cf922d Passing use_auth = True and also having different checks which is missed by the is_vision function 2026-02-19 02:55:46 +00:00
Roland Tannous
c876b38780 reduce dataset_num_proc to 1/4 of cpu_count 2026-02-18 20:53:32 +00:00
Roland Tannous
648b29ac9b Merge pull request #155 from unslothai/fix/sft-tokenizer-unwrap-for-vlm-text 
fix: Unwrap ProcessorMixin to raw tokenizer for text-only SFTTrainer on VLM-architecture models
 2026-02-18 23:19:57 +04:00
Roland Tannous
23214c41c0 fix: unwrap ProcessorMixin to raw tokenizer for text-only SFTTrainer on VLM-architecture models 2026-02-18 19:13:20 +00:00
Roland Tannous
9840864662 Merge remote-tracking branch 'origin/nightly' into fix/dataset-mapping-vlm-text-datasets 2026-02-18 18:11:45 +04:00
Roland Tannous
ad638118b9 Merge remote-tracking branch 'origin/nightly' into feature/colab-notebook 2026-02-18 17:41:07 +04:00
Roland Tannous
ae040cf681 renamed UNSLOTH_FLEX_ATTENTION to UNSLOTH_ENABLE_FLEX_ATTENTION 2026-02-18 09:21:53 +00:00
Roland Tannous
e3f4a9eb32 Disable flex attention on Blackwell+ GPUs (sm_120+) at startup 2026-02-18 08:58:25 +00:00
Roland Tannous
29b25169c0 Scale dataset num_proc dynamically to cpu_count//3 instead of hardcap 8 2026-02-18 08:38:53 +00:00
Manan17
949e57c334 fixing the hangup of training after multiple back to back training processes 2026-02-18 08:18:13 +00:00
Manan17
db0fa1a270 Dividing the total cpu_count // 3 2026-02-18 07:59:57 +00:00
Roland Tannous
c616697b22 Merge pull request #148 from unslothai/fix/linear 
linear fix
 2026-02-18 11:10:16 +04:00
Manan17
c832c903b4 fix the linear path on backend 2026-02-18 07:08:32 +00:00
Roland Tannous
38e577ef26 Merge pull request #145 from unslothai/fix/check-format-sample 
fix: stream HF datasets in check-format endpoint to avoid full d…
 2026-02-18 10:49:17 +04:00
Roland Tannous
9d737559a5 debug statements 2026-02-18 00:37:09 +00:00
Roland Tannous
d94a1e0289 fix: normalize target_modules [all-linear] list to string for Unsloth/PEFT compatibility 2026-02-18 00:32:21 +00:00
Roland Tannous
f0613f5d07 fix: defensively rename VLM chat column to match model's forward() signature 2026-02-17 23:49:13 +00:00
Roland Tannous
c0f210bc2a fix: cap dataset.map() num_proc to 8 to prevent CUDA fork deadlocks 2026-02-17 23:12:45 +00:00
Roland Tannous
4d0c6d20b3 fix: fix: stream HF datasets in check-format endpoint to avoid full downloads; add info logging to model config endpoints 2026-02-17 22:53:29 +00:00
Shine1i
b31461790f feat: enhance training stop and reset flow with detailed checks 2026-02-17 23:32:22 +01:00
Leo Borcherding
84b9a8aef6 Merge nightly into feature/colab-notebook - resolved setup.sh conflicts 2026-02-17 15:51:14 -06:00