Repository navigation

All pull requests

Labels (27)Milestones (3)New pull request

Open

(65)

Closed

(551)

Expose skipped documents in downstream evaluation metrics
#921dajiaohuang opened Oct 23, 2026
docs: add Simplified Chinese translation (README.zh.md)
#920JasonYeYuhe opened Oct 12, 2026
4
Retry transient network errors in _http_file_size
#918pujitha24 opened Sep 17, 2026
1
Fix FSDP parameter counts and HF generation compatibility
#916avinashgola opened Sep 6, 2026
Replace black/isort with ruff format
#909korbonits opened Apr 11, 20263 tasks done
Improve run_dataloader.py usability and performance feedback(addresses #869)
#902ezylopx5 opened Jan 14, 2026
Bump actions/upload-artifact from 4 to 5
#893dependabot[bot] opened Oct 24, 2025Bot
1
Bump softprops/action-gh-release from 1 to 2
#892dependabot[bot] opened Oct 24, 2025Bot
Bump actions/download-artifact from 4 to 6
#891dependabot[bot] opened Oct 24, 2025Bot
Bump actions/checkout from 3 to 5
#890dependabot[bot] opened Oct 24, 2025Bot
Bump actions/cache from 3 to 4
#889dependabot[bot] opened Oct 24, 2025Bot
Bump torch from 2.0.1 to 2.8.0 in /inference/efficiency/dependencies/previous_version
#888dependabot[bot] opened Oct 23, 2025Bot
Add MuonW optimizer: Muon with AdamW fallback for non-matrix parameters
#882JenWei0312 opened Aug 31, 2025
2
support "upload" to local files
#881canonic-epicure opened Aug 28, 2025
1
Update README.md for dependencies installation
#872tigermlt opened Aug 5, 2025
docs: fix minor typo in README
#833koshieguchi opened May 5, 2025
Fix tokenizer.encode() to respect add_special_tokens=False parameter
#821ved1beta opened Apr 7, 2025Contributor
1
Fix LayerNorm, Multi-Query Attention, and Weight-Tying in OLMo to HF Conversion Script
#820ved1beta opened Apr 3, 2025Contributor
[Feature] add tensorbard support to the training framework
#811lyuwen opened Mar 21, 2025
5
Updated memmap_dtype to automatically choose based on vocab size
#795aman-17 opened Feb 12, 2025Contributor
2
Decoupled Momentum Optimization
#771peter-sk opened Dec 24, 2024
6
update mean reduction zloss to ignore labels == ignore_index
#760jasonkrone opened Dec 4, 2024
1
Adds support for converting from safetensors
#740soldni opened Oct 23, 2024Contributor
1
Create an eval-only script for existing ckpts
#736liujch1998 opened Oct 20, 2024Contributor
17
Add regression tests for training
#7302015aroras opened Oct 7, 2024Collaborator
1

Navigation Menu

Repository navigation

All pull requests

Expose skipped documents in downstream evaluation metrics

docs: add Simplified Chinese translation (README.zh.md)

Retry transient network errors in _http_file_size

Fix FSDP parameter counts and HF generation compatibility

Replace black/isort with ruff format

Improve run_dataloader.py usability and performance feedback(addresses #869)

Bump actions/upload-artifact from 4 to 5

Bump softprops/action-gh-release from 1 to 2

Bump actions/download-artifact from 4 to 6

Bump actions/checkout from 3 to 5

Bump actions/cache from 3 to 4

Bump torch from 2.0.1 to 2.8.0 in /inference/efficiency/dependencies/previous_version

Add MuonW optimizer: Muon with AdamW fallback for non-matrix parameters

support "upload" to local files

Update README.md for dependencies installation

docs: fix minor typo in `README`

Fix tokenizer.encode() to respect add_special_tokens=False parameter

Fix LayerNorm, Multi-Query Attention, and Weight-Tying in OLMo to HF Conversion Script

[Feature] add tensorbard support to the training framework

Updated memmap_dtype to automatically choose based on vocab size

Decoupled Momentum Optimization

update mean reduction zloss to ignore labels == ignore_index

Adds support for converting from safetensors

Create an eval-only script for existing ckpts

Add regression tests for training

Pull requests