Skip to content

Pull requests: marin-community/levanter

New pull request New
25 Open 947 Closed
25 Open 947 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Update transformers requirement from <5.0,>=4.57.1 to >=4.57.1,<6.0 dependencies Pull requests that update a dependency file python Pull requests that update python code
#1298 opened Jan 26, 2026 by dependabot bot Loading…
Update gcsfs requirement from <2026,>=2024.2 to >=2024.2,<2027 dependencies Pull requests that update a dependency file python Pull requests that update python code
#1297 opened Jan 12, 2026 by dependabot bot Loading…
Update fsspec[http] requirement from <2026,>=2024.2 to >=2024.2,<2027 dependencies Pull requests that update a dependency file python Pull requests that update python code
#1296 opened Jan 12, 2026 by dependabot bot Loading…
Update safetensors[numpy] requirement from <0.7.0,>=0.4.2 to >=0.4.2,<0.8.0 dependencies Pull requests that update a dependency file python Pull requests that update python code
#1295 opened Nov 24, 2025 by dependabot bot Loading…
Flash Gated Delta Rule kernels in Pallas
#1292 opened Nov 4, 2025 by Calvin-Xu • Draft
@Calvin-Xu
Fix torch test meshes, run torch tests in CI
#1289 opened Nov 3, 2025 by dlwh Loading…
4
expand relative paths to full paths for the cache builder
#1274 opened Oct 22, 2025 by dlwh Loading…
1
6
Support syncing weights from Levanter to vLLM
#1272 opened Oct 20, 2025 by AlienKevin • Draft
5
Dataloading should fail if requested Mixture is not feasible
#1248 opened Oct 10, 2025 by ahmeda14960 Loading…
4
MuP In Levanter
#1231 opened Oct 3, 2025 by Helw150 • Draft
20
WIP: Add style prefix token utilities
#1228 opened Oct 3, 2025 by chris544460 • Draft
Add async request queue processing to inference engine codex
#1211 opened Sep 29, 2025 by dlwh • Draft
Add debug comparison between TPU and reference ragged attention codex
#1190 opened Sep 25, 2025 by dlwh Loading…
Add smuggler utilities for JAX transforms codex
#1167 opened Sep 17, 2025 by dlwh • Draft
2
feat(loss): add pallas kernels for fused cross-entropy loss
#1166 opened Sep 17, 2025 by BrendanGraham14 Loading…
4
WIP: Add JIT-ed UL2R functions
#1158 opened Sep 12, 2025 by jyc • Draft
7 tasks
22
Fix: Correct implementation in scion.py
#1146 opened Sep 5, 2025 by Wanyun0327 Loading…
4
docs: outline unified checkpoint strategy codex
#1136 opened Aug 15, 2025 by dlwh • Draft
Add multi-book sliding evaluation codex
#1121 opened Aug 5, 2025 by ahmeda14960 Loading…
Tree-of-actors implementation of Ray TPU scheduling
#1054 opened Jul 11, 2025 by yifanmai • Draft
1
Adding PSGD-QUAD to optimizer collection
#1022 opened Jul 1, 2025 by evanatyourservice Loading…
4
Add dead neuron histograms codex
#1021 opened Jul 1, 2025 by dlwh Loading…
Support parameter tags in optimizer weight decay codex
#996 opened Jun 17, 2025 by dlwh Loading…
1
UpcycleLm script
#987 opened Jun 12, 2025 by blahBlahhhJ Loading…
3
Allow for tokenizers/preprocessors to change batch size
#866 opened Jan 24, 2025 by Aphoh Loading…
6
ProTip! Exclude everything labeled bug with -label:bug.