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

783 commits

This branch
This branch
All branches
Author SHA1 Message Date
Manan17
4faa6e9152 fixing sesame model 2026-03-07 19:06:00 +00:00
Manan17
5d1a162ddd remove tracked OuteTTS embedded repo reference 2026-03-05 08:44:23 +00:00
Manan17
8203637d89 resolved merge conflicts 2026-03-05 07:59:43 +00:00
Manan17
2fa933640e fix SNAC training crash on variable-length sequences with DataCollatorForSeq2Seq 2026-03-05 07:04:53 +00:00
Roland Tannous
37bff450d5 Merge pull request #314 from unslothai/fix/vlm-dataset-conversion-error-handling-local 
Fix VLM training abort on URL-based dataset conversion failure
 2026-03-05 10:10:58 +04:00
Roland Tannous
a1706c894f fix: check for http(s) prefix instead of bare string type for URL detection 2026-03-05 06:10:10 +00:00
Roland Tannous
2116cc5cca fix: remove benchmark scripts from git tracking 
These are standalone benchmark scripts that were force-added despite being
gitignored. They have no test functions and run network calls at module
level, which breaks pytest collection in CI.
 2026-03-05 06:06:47 +00:00
Roland Tannous
9907afac9f Merge pull request #313 from unslothai/feature/index-range-dataset-slicing 
Fix: clear dataset slice state on file upload
 2026-03-05 03:48:10 +04:00
Roland Tannous
b38df21bb5 fix: clear dataset slice state when switching to uploaded file 
Prevents stale slice values from silently truncating uploaded datasets.
 2026-03-04 23:42:23 +00:00
Roland Tannous
e04b9d53d6 feat: parallel URL image probe with time estimate and progress reporting 
- Add 200-sample parallel probe using ThreadPoolExecutor + safe_num_proc
  to estimate download speed and failure rate before full conversion
- Abort with clear error if >=30% of probe images fail to download
- Show estimated download time in the training overlay modal
- Parallel batch conversion for URL-based datasets (vs sequential for local)
- Add warning field to /check-format response for URL-based image datasets
- Display URL warning in dataset preview dialog (amber banner)
- Thread progress_callback from trainer through format_and_template_dataset
  to convert_to_vlm_format for real-time status updates
 2026-03-04 23:40:38 +00:00
Roland Tannous
db30b4105f test: add parallel download benchmark with ThreadPoolExecutor 2026-03-04 23:29:43 +00:00
Roland Tannous
1f03754c95 feat: add tqdm progress bar to VLM conversion and download benchmark test 2026-03-04 23:29:43 +00:00
Roland Tannous
63f723cc36 fix: add early probe to fail fast on datasets with too many broken image URLs 2026-03-04 23:29:43 +00:00
Roland Tannous
9487d17b94 fix: use fsspec for URL image downloads with per-sample error handling 2026-03-04 23:29:43 +00:00
Roland Tannous
e6eea64df5 test: add URL image loading comparison script 2026-03-04 23:29:43 +00:00
Roland Tannous
b376c54c21 fix: abort training pipeline on dataset conversion failure 2026-03-04 23:29:43 +00:00
Roland Tannous
bc244aeb23 fix: cast URL image columns to HF Image() type in VLM conversion 2026-03-04 23:29:43 +00:00
Roland Tannous
70434210c7 Merge pull request #312 from unslothai/feature/index-range-dataset-slicing 
Add index range dataset slicing to Studio training page
 2026-03-05 03:25:21 +04:00
Roland Tannous
945cfa9460 fix: remove unnecessary tooltip copy from train split start 2026-03-04 23:24:09 +00:00
Roland Tannous
263cd19c44 refactor: move train split slice controls back to Advanced section 
Place Train Split Start / End inputs inside the Advanced collapsible
with descriptive tooltips clarifying they slice the training split.
Revert the selectors component to its original eval-split-only layout.
 2026-03-04 23:24:09 +00:00
Roland Tannous
44905e9216 refactor: move index range fields next to eval split in 3-col grid 
Place Slice Start and Slice End inputs alongside the Eval Split
selector in a single row (grid-cols-3) so the dataset card stays
compact. Remove the duplicate controls from the Advanced section.
 2026-03-04 23:24:09 +00:00
Roland Tannous
64889cd5fc feat: add index range dataset slicing to studio training page 
Add Start/End index inputs under Advanced in the dataset card,
allowing users to slice a dataset by row range before training.
Wired end-to-end: frontend store, API payload, backend Pydantic
model, and trainer dataset loading (inclusive on both ends).
 2026-03-04 23:24:09 +00:00
Roland Tannous
52913a4c69 Merge pull request #311 from unslothai/revert-310-feature/index-range-dataset-slicing 
Revert "Add index range dataset slicing to Studio training page"
 2026-03-05 03:23:31 +04:00
Roland Tannous
9333f99dd3 Revert "Add index range dataset slicing to Studio training page" 2026-03-05 03:21:07 +04:00
Roland Tannous
40113f2f9c Merge pull request #310 from unslothai/feature/index-range-dataset-slicing 
Add index range dataset slicing to Studio training page
 2026-03-05 03:20:31 +04:00
Roland Tannous
43feb6c2e2 fix: remove unnecessary tooltip copy from train split start 2026-03-04 23:15:49 +00:00
Roland Tannous
ed702b4850 refactor: move train split slice controls back to Advanced section 
Place Train Split Start / End inputs inside the Advanced collapsible
with descriptive tooltips clarifying they slice the training split.
Revert the selectors component to its original eval-split-only layout.
 2026-03-04 23:07:36 +00:00
Roland Tannous
8d9f195ec0 refactor: move index range fields next to eval split in 3-col grid 
Place Slice Start and Slice End inputs alongside the Eval Split
selector in a single row (grid-cols-3) so the dataset card stays
compact. Remove the duplicate controls from the Advanced section.
 2026-03-04 22:35:17 +00:00
Roland Tannous
02b17ec6d9 feat: add index range dataset slicing to studio training page 
Add Start/End index inputs under Advanced in the dataset card,
allowing users to slice a dataset by row range before training.
Wired end-to-end: frontend store, API payload, backend Pydantic
model, and trainer dataset loading (inclusive on both ends).
 2026-03-04 21:48:40 +00:00
Roland Tannous
760f4a3609 test: add parallel download benchmark with ThreadPoolExecutor 2026-03-04 14:30:11 +00:00
Roland Tannous
cc11f066b1 feat: add tqdm progress bar to VLM conversion and download benchmark test 2026-03-04 13:30:27 +00:00
Manan Shah
a8fe6c28a3 Delete studio/TESTING.md 2026-03-04 03:47:19 -07:00
Manan17
ee23628b0c dynamic detection of audio models and fixing autoencoder issues 2026-03-04 10:44:44 +00:00
Roland Tannous
7804a4db2e fix: add early probe to fail fast on datasets with too many broken image URLs 2026-03-04 08:05:40 +00:00
Roland Tannous
9cbd3d44a7 fix: use fsspec for URL image downloads with per-sample error handling 2026-03-04 07:50:55 +00:00
Roland Tannous
060e2e1abf test: add URL image loading comparison script 2026-03-04 07:39:35 +00:00
Roland Tannous
c6d82a6fad fix: abort training pipeline on dataset conversion failure 2026-03-04 06:42:48 +00:00
Roland Tannous
3c4bf80cc2 fix: cast URL image columns to HF Image() type in VLM conversion 2026-03-04 06:42:37 +00:00
Roland Tannous
aa921dee76 Merge pull request #305 from unslothai/fix/dropdown-layout-shift 
Prevent select dropdowns from shifting layout when opened
 2026-03-04 10:19:27 +04:00
Roland Tannous
874e321fdf Merge pull request #306 from unslothai/fix/hf-dataset-error-message 
Sanitize dataset script errors and persist training start error
 2026-03-04 10:14:26 +04:00
Roland Tannous
84ad61bfa2 Remove overly broad .py check from dataset error normalization 2026-03-04 06:13:47 +00:00
Roland Tannous
14a8fbc650 Merge pull request #287 from unslothai/fix/duplicate-def-inference 
Deleted duplicate definitions for load_for_eval, load_adapter, and load_model_simple in core Inference
 2026-03-04 10:06:04 +04:00
Roland Tannous
e8dfa8446d Merge pull request #289 from unslothai/fix/datasets-auth 
Added auth to dataset endpoints
 2026-03-04 08:21:42 +04:00
imagineer99
7bc5bc26f6 fix: sanitize dataset script errors and persist training start error 2026-03-03 20:15:23 +00:00
Roland Tannous
9dbee02417 Updated README 2026-03-03 18:42:35 +00:00
imagineer99
981b618c1c fix: prevent select scroll-lock margin from shifting layout 2026-03-03 18:37:46 +00:00
Roland Tannous
c89c3e79be Merge pull request #296 from unslothai/feature/windows-native-support 
PR: Windows Native Support + llama.cpp Build Migration
 2026-03-03 22:23:35 +04:00
Roland Tannous
0b4f6a1343 Merge branch 'nightly' into feature/windows-native-support 2026-03-03 22:23:18 +04:00
Roland Tannous
8c333e2b81 chore: add cross-platform Python installer with updated unsloth patch URLs 2026-03-03 17:31:59 +00:00
Roland Tannous
5f98d232d0 fix: align llama-server binary discovery with upstream unsloth-zoo paths 2026-03-03 17:03:01 +00:00