Skip to content

Pull requests: vllm-project/tpu-inference

New pull request New
46 Open 911 Closed
46 Open 911 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

[GPT-OSS] add unit test for GPT-OSS
#978 opened Oct 30, 2025 by bzgoogle • Draft
1
[DeepSeek][Do-Not-Merge] experimental PR for DeepSeek optimization
#977 opened Oct 30, 2025 by bzgoogle • Draft
1
Fix kv cache shape for head_dim=64
#976 opened Oct 30, 2025 by yaochengji Loading…
1
[Model] Add vision encoder and input embeddings merger warmup for Qwen2.5 VL model
#972 opened Oct 29, 2025 by KWang1998 Loading…
5
[FIX] Add dummy get_input_embeddings to fix vLLM model type check
#971 opened Oct 29, 2025 by kuafou Loading…
1
[TPU host offload][FIX] Keep KV cache using NamedSharding on CPU
#970 opened Oct 29, 2025 by juncgu-google Loading…
1
[Llama4/JAX] Llama4 FP8 Quantized Weight Loading and Sharding
#962 opened Oct 28, 2025 by sierraisland Loading…
37
[Spec Decoding] Reduce TPU <-> CPU data transfer
#961 opened Oct 28, 2025 by Lumosis Loading…
1
Update README.md
#956 opened Oct 27, 2025 by bvrockwell Loading…
1
Install runai-model-streamer module in Dockerfile
#955 opened Oct 27, 2025 by amacaskill Loading…
dtype in ModelConfig will be implicitly casted to torch.dtype so in tpu_jax, we need to check for torch dtype as well
#945 opened Oct 27, 2025 by lc5211 Loading…
1
[GPT-OSS] enable attention-sink
#943 opened Oct 27, 2025 by bzgoogle • Draft
1
[do not review yet] Get rid of a2a
#939 opened Oct 24, 2025 by vanbasten23 • Draft
1
[multi-host] add quick start guide
#928 opened Oct 23, 2025 by Lumosis Loading…
1
[Requirements] Bump JAX/JAXLib to 0.8.0
#927 opened Oct 23, 2025 by jrplatin Loading…
1
[Kernel] Refactor ragged_paged_attention to proxy for default and hd64
#918 opened Oct 22, 2025 by yaochengji • Draft
1
[Requirements] Bump TPU Info to 0.6.0
#917 opened Oct 22, 2025 by jrplatin Loading…
2
PP for single host
#914 opened Oct 21, 2025 by Chenyaaang • Draft
[WIP] Add Qwen3-Omni model
#896 opened Oct 19, 2025 by eitanporat Loading…
add jax support for Qwen2VL
#893 opened Oct 18, 2025 by shungcp Loading…
[Doc] Docker guide extended
#890 opened Oct 17, 2025 by hosseinsarshar Loading…
Data Parallelism support
#865 opened Oct 14, 2025 by wenxindongwork Loading…
@wenxindongwork
22
[CI] remove lora_bias_stacked as it is deprecated in vllm
#835 opened Oct 11, 2025 by bzgoogle Loading…
3
lora spmd
#802 opened Oct 8, 2025 by vanbasten23 • Draft
feat(ci): Record vLLM and tpu-inference commit hashes and refactor commit logic
#795 opened Oct 7, 2025 by dennisYehCienet Loading…
17
ProTip! no:milestone will show everything without a milestone.