Skip to content

Pull requests: keras-team/keras-hub

New pull request New
74 Open 1,788 Closed
74 Open 1,788 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

Fix attention score returns in CachedMultiHeadAttention
#2799 opened Jul 9, 2026 by GargiGupta-io Loading…
6 tasks done
1
12
Add model-parallel layout maps for 14 GQA model families + fix key/value sharding crashes (incl. Gemma v1) Gemma Gemma model specific issues keras-team-review-pending
#2795 opened Jul 8, 2026 by pctablet505 Collaborator Loading…
29
Serve KerasHub CausalLM models through vLLM on TPU Gemma Gemma model specific issues
#2794 opened Jul 7, 2026 by anthony-etim Loading…
6 tasks done
59
Prevent division by zero in anchor normalizer calculation
#2791 opened Jul 6, 2026 by pzarzycki Loading…
6 tasks done
10
Add backend-native safetensor loading for faster conversion
#2790 opened Jul 4, 2026 by bzantium Loading…
1
Enable LiteRT export tests for the TensorFlow backend Gemma Gemma model specific issues
#2782 opened Jun 27, 2026 by pctablet505 Collaborator • Draft
4
[Gemma3] Reduce concatenate and copy in KV cache update to improve perf Gemma Gemma model specific issues keras-team-review-pending
#2781 opened Jun 25, 2026 by gaurides Contributor Loading…
2 tasks
7
Fix sentencepiece special tokens Gemma Gemma model specific issues
#2779 opened Jun 25, 2026 by Nikhilyammani007 Loading…
6 tasks done
12
Fix RetinaNet LiteRT export test: assert on real output keys
#2778 opened Jun 25, 2026 by pctablet505 Collaborator • Draft
2
Fix D-Fine torch/LiteRT export: static shapes in split + deformable attention
#2777 opened Jun 25, 2026 by pctablet505 Collaborator • Draft
2
Fix DeepLabV3 & SegFormer inference/LiteRT export (Resizing training leak)
#2776 opened Jun 25, 2026 by pctablet505 Collaborator • Draft
3
Fix LoRA save/load to use stable layer paths (#2701) Gemma Gemma model specific issues stat:awaiting response from contributor
#2774 opened Jun 23, 2026 by ManojPawar2 Loading…
1
7
Fix Gemma3/3n/4 preprocessor crashes with Python-only tokenizer Gemma Gemma model specific issues
#2773 opened Jun 23, 2026 by pctablet505 Collaborator Loading…
7
Fix Adam optimizer underflow for LoRA in bfloat16 Gemma Gemma model specific issues
#2772 opened Jun 22, 2026 by ManojPawar2 Loading…
4
Add safe_mode guard to PARSeq and Whisper tokenizer file loading
#2771 opened Jun 19, 2026 by a1ohadance Loading…
1
fix docstrings and add warning and validation for max_length when preprocessor is None
#2765 opened Jun 12, 2026 by maitry63 Contributor Loading…
6 tasks done
1
10
Add Gemma 4 12B (Unified) encoder-free multimodal support to hub Gemma Gemma model specific issues
#2758 opened Jun 5, 2026 by laxmareddyp Collaborator • Draft
6 tasks done
69
Improve validation for ByteTokenizer.sequence_length
#2756 opened Jun 4, 2026 by maitry63 Contributor Loading…
6 tasks done
1
5
Fix LoRA save/load to use stable path-based keys Gemma Gemma model specific issues
#2725 opened May 7, 2026 by pctablet505 Collaborator • Draft
1
6
Add TIPSv2 vision-language dual-encoder model to hub new model For PRs that contribute a new model to the Keras Hub registry.
#2721 opened May 5, 2026 by laxmareddyp Collaborator • Draft
6 tasks done
6
Add OpenAI Privacy Filter Model to Keras-Hub new model For PRs that contribute a new model to the Keras Hub registry.
#2718 opened May 3, 2026 by laxmareddyp Collaborator Loading…
6 tasks done
6
Add SmolVLM2 multimodal vision-language model new model For PRs that contribute a new model to the Keras Hub registry.
#2717 opened May 3, 2026 by laxmareddyp Collaborator Loading…
6 tasks done
7
Implement KerasHub → LiteRT-LM export with prefill/decode signatures Gemma Gemma model specific issues
#2705 opened Apr 25, 2026 by pctablet505 Collaborator • Draft
25
Add blip2 new model For PRs that contribute a new model to the Keras Hub registry. stat:awaiting response from contributor
#2699 opened Apr 24, 2026 by yassienashrafwasfy Loading…
1
92
Enable LiteRT export tests on the torch backend Gemma Gemma model specific issues
#2698 opened Apr 23, 2026 by pctablet505 Collaborator Loading…
6
ProTip! Updated in the last three days: updated:>2026-07-12.