Skip to content

Pull requests: flashinfer-ai/flashinfer

New pull request New
68 Open 1,570 Closed
68 Open 1,570 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

Super tiny fix version
#2199 opened Dec 11, 2025 by fzyzcjy Loading…
5 tasks
4
docs: Document CUDA version support in README and installation page
#2197 opened Dec 10, 2025 by bkryu Loading…
5 tasks done
2025 Dec
@bkryu
6
docs: Fix inaccurate API docstrings for attention prefill
#2196 opened Dec 10, 2025 by bkryu Loading…
5 tasks
2025 Dec
@bkryu
4
Permute page table in benchmarking
#2194 opened Dec 10, 2025 by jhjpark Loading…
3 of 5 tasks
6
feat: unit-test and api change, w4a8 grouped-gemm fused MoE for SM90
#2193 opened Dec 10, 2025 by jimmyzho Loading…
5 tasks
@jimmyzho
14
refactor: update fa3 codebase [part 2]
#2192 opened Dec 9, 2025 by yzh119 Loading…
4 of 5 tasks
13
Fix for moe on sm110
#2190 opened Dec 9, 2025 by jhalabi-nv Loading…
3 of 5 tasks
10
chore: Update CODEOWNERS automated maintenance
#2186 opened Dec 8, 2025 by flashinfer-bot Loading…
4
Add CUDA graph buffers for persistent attention
#2185 opened Dec 7, 2025 by Edenzzzz Loading…
5 tasks
13
Fix/moe_sm110 (to be tested)
#2183 opened Dec 6, 2025 by aleozlx • Draft
5 tasks
4
[Flashinfer-Bench integration] HF end-to-end inference
#2151 opened Nov 30, 2025 by sfc-gh-goliaro • Draft
5 tasks
5
Enable Hopper FA3 FP8 attention
#2148 opened Nov 28, 2025 by nvpohanh • Draft
5 tasks
2
make DeepGEMM swapAB available for linear gemm SM90
#2131 opened Nov 22, 2025 by katec846 Loading…
3 of 5 tasks
17
A unified API for the MNNVL and single-node AllReduce kernels.
#2130 opened Nov 21, 2025 by nvmbreughe • Draft
5 tasks
9
feat: support variable sequence length in decode kernel of trtllm-gen attention
#2125 opened Nov 20, 2025 by yaoyaoding Loading…
4 of 5 tasks
1
17
perf: using multi-cta optimization for top-k/top-p
#2119 opened Nov 20, 2025 by yzh119 Loading…
4 of 5 tasks
9
Refactor trtllm_mnnvl_allreduce
#2118 opened Nov 20, 2025 by timlee0212 Loading…
5 tasks done
78
feat: support more head dim in RoPE kernel
#2109 opened Nov 19, 2025 by raayandhar Loading…
5 tasks done
25
Port TRT-LLM communication kernels to flashinfer
#2102 opened Nov 18, 2025 by djns99 Loading…
5 tasks done
82
make DeepGEMM swapAB available for linear gemm SM90
#2101 opened Nov 17, 2025 by xuanzic Loading…
5 tasks
9
feat: add sink to flashinfer decode
#2087 opened Nov 13, 2025 by djmmoss Loading…
14
feat: BF16 GEMM using CUTLASS backend for SM100
#2070 opened Nov 10, 2025 by raayandhar Loading…
5 tasks done
47
Rebase FP8 SM100 Cutlass FMHA Attention to main (original PR#1238)
#2047 opened Nov 5, 2025 by pavanimajety Loading…
5 tasks
6
Added an initial implementation of Q and KV Cache in fp8 and to use t…
#2035 opened Nov 4, 2025 by Anerudhan • Draft
3 tasks done
1
Refactor flashinfer/__init__.py so that applications could selectively pack submodules without modifying __init__.py
#2027 opened Nov 3, 2025 by bangshengtang Loading…
5 tasks done
9
ProTip! Follow long discussions with comments:>50.