-
Notifications
You must be signed in to change notification settings - Fork 124
Pull requests: ovg-project/kvcached
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(vllm): preserve duplicate prefix cache blocks
#402
opened Jul 22, 2026 by
shipiyouniao
Contributor
Loading…
feat(sglang): pool DeepSeek-V4 SWA KV cache
#400
opened Jul 21, 2026 by
shipiyouniao
Contributor
•
Draft
feat(vllm): support contiguous KV layout for hybrid linear-attention (Qwen3-Next/GDN)
#397
opened Jul 20, 2026 by
RixinLiu
Collaborator
Loading…
ci: fix integration branch pre-commit checks
#393
opened Jul 19, 2026 by
shipiyouniao
Contributor
Loading…
fix(vllm): evict prefix-cache blocks by page to release memory (#359)
#390
opened Jul 17, 2026 by
ryanManocha
Loading…
fix(sglang): use storage dtype for FP8 KV pools
#388
opened Jul 16, 2026 by
shipiyouniao
Contributor
Loading…
Fix vLLM Gemma4 support by adding v2 module fallback and patch target fallback
#386
opened Jul 14, 2026 by
mahendrarathore1742
Contributor
Loading…
fix(sglang): derive free page indices on CPU
#384
opened Jul 5, 2026 by
shipiyouniao
Contributor
Loading…
fix: make FTensor page map operations idempotent
#383
opened Jul 5, 2026 by
shipiyouniao
Contributor
Loading…
fix(vllm): floor max cached blocks at 1 for sub-block-size budgets (#343)
#380
opened Jul 2, 2026 by
SuperMarioYL
Loading…
Fix unsigned underflow in available physical page calculation
#377
opened Jul 2, 2026 by
cmssmc00
Loading…
feat(sglang): account DeepSeek-V4 runtime reservations
#376
opened Jul 1, 2026 by
shipiyouniao
Contributor
Loading…
fix: propagate TP and PP KV mapping failures
#373
opened Jul 1, 2026 by
shipiyouniao
Contributor
Loading…
docs: Add comprehensive bilingual documentation with MkDocs site
#368
opened Jun 29, 2026 by
tianhao909
Loading…
6 tasks done
fix: roll back partial KV allocation on page miss
#365
opened Jun 24, 2026 by
shipiyouniao
Contributor
Loading…
feat: support uniform cross-layer KV cache layout for CPU offloading
#361
opened Jun 15, 2026 by
lianghao208
Contributor
Loading…
fix map_to_kv_tensors/unmap_from_kv_tensors in MLA non-contiguous layout
enhancement
New feature or request
#357
opened Jun 9, 2026 by
lianghao208
Contributor
Loading…
fix: defer tp prealloc until first alloc
#341
opened May 21, 2026 by
shipiyouniao
Contributor
Loading…
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.