Skip to content

Pull requests: google-deepmind/gemma

New pull request New
156 Open 331 Closed
156 Open 331 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

feat: add rolling cache turn pruning for ChatSampler (Issue #675)
#703 opened Jun 23, 2026 by Paramveersingh-S Loading…
1
Fix Orbax->ROC conversion for Gemma 4 MoE checkpoints.
#702 opened Jun 22, 2026 by copybara-service Bot Loading…
1
Fix: Rolling KV cache and top-k logit trimming (fixes #675)
#698 opened Jun 14, 2026 by medmomoait Loading…
1
1
fix(t5gemma): add missing f-prefix in _normalize_token error message
#697 opened Jun 14, 2026 by Osamaali313 Loading…
2
Add sharding support to ChatSampler.
#696 opened Jun 12, 2026 by copybara-service Bot Loading…
Test
#693 opened Jun 11, 2026 by copybara-service Bot Loading…
Fix: Resolve KV Cache OOM & Logit Extraction Memory Spikes
#678 opened Jun 7, 2026 by prince-shakyaa Loading…
1
1
Support loading checkpoints from ocp.PyTreeCheckpointer format.
#673 opened Jun 5, 2026 by copybara-service Bot Loading…
fix: use f-string in _normalize_token error messages
#669 opened Jun 3, 2026 by moltate Loading…
1
1
fix(gqa): replace float division with integer division in GQA reshape
#668 opened May 29, 2026 by Sumu004 Loading…
1
Flip default freeze to False in konfig.resolve and remove ImmutableDict.
#664 opened May 28, 2026 by copybara-service Bot Loading…
1
Fix MoERagged docstring return reference
#663 opened May 26, 2026 by dtype2100 Loading…
1
fix(utils): download tokenizer models locally via HTTP to avoid TF GCS C++ segfault on macOS
#661 opened May 23, 2026 by prince-shakyaa Loading…
1
1
fix: np.float128 platform crash on macOS/Windows & missing f-string in error message
#659 opened May 23, 2026 by prince-shakyaa Loading…
6 tasks done
2
4
Fix typos in LoRA einsum utils docstring
#655 opened May 22, 2026 by AshNicolus Loading…
Add FORBIDDEN_TOKENS to Gemma4Tokenizer covering image + audio placeholders
#649 opened May 16, 2026 by dparikh79 Loading…
4 tasks done
1
Update Gemma 4 status in README
#645 opened May 11, 2026 by Rhabibi1609 Loading…
2
fix: use integer division (//) in GQA reshape operations
#643 opened May 4, 2026 by tomohiro86 Loading…
3 tasks
1
Optionally allow models to keep the last cache item at prefill time.
#636 opened Apr 28, 2026 by copybara-service Bot Loading…
1
Fix typo in LoRA fine-tuning comment
#632 opened Apr 26, 2026 by syrkis Loading…
1
Implement Prefix Tuning for Gemma models.
#631 opened Apr 22, 2026 by copybara-service Bot Loading…
1
Fix jaxtyping shape mismatch for multimodal inputs in Gemma3nTransformer
#624 opened Apr 17, 2026 by YADAV1825 Loading…
1
docs: Add Python 3.11+ requirement
#623 opened Apr 13, 2026 by DJH52 Loading…
1
fix: two bugs in Sampler — missing f-string in _normalize_token and XOR logic error in _remove_eos_token
#618 opened Apr 7, 2026 by Dhakshin2007 Loading…
fix: correct misleading ValueError message in Attention._call_
#617 opened Apr 7, 2026 by Dhakshin2007 Loading…
ProTip! What’s not been updated in a month: updated:<2026-05-23.