Repository navigation

All pull requests

Labels (133)Milestones (10)New pull request

[CI] [GHA] Skip rerunner analysis when the run has no logs
#39368blairducrayoppat opened Oct 30, 2026
[GPU] Add MoE int3 dpas kernel
#39366NinaWie opened Oct 30, 2026Contributor
[core] Roll back partially written entries in SingleFileStorage
#39365vmanvs opened Oct 30, 2026Contributor
[NPUW] PA front-end: get rid of per-sequence loop
#39362parthmah14 opened Oct 30, 2026Contributor
[Arm CPU] [Torch.compile Path] Support for vllm dynamic-4-bit kernel
#39361abhijain1204fujitsu opened Oct 30, 2026Contributor
[PyOV] Update documentation with Python 3.15 support
#39360p-wysocki opened Oct 30, 2026Contributor 2026.5
1
fix: cover YaRN RoPE in IncreasePositionIdsPrecision
#39359utsab345 opened Oct 30, 2026
1
[CPU][ARM] KleidiAI INT8 FC: select the GEMV for M == 1 and the 16x4 GEMM for M > 1 on dotprod and i8mm cores
#39358clemlesne opened Oct 30, 2026
2
[GPU] Fix zero shape RemoteTensor as input and fullyconnected execution
#39356ZackyLake opened Oct 30, 2026Contributor
6
[GPU] Add llama.cpp-specific in-place recurrent cache update ops
#39355wine99 opened Oct 30, 2026Contributor
1
[GPU] Defer dynamic output argument binding until execution preparation
#39354wine99 opened Oct 30, 2026Contributor
1
[CVS-193582][NPUW][WIP] add qwen3guard-stream-0.6B support on NPUW
#39353bopeng1234 opened Oct 30, 2026Contributor
3
Bump the github-actions group with 3 updates
#39352dependabot[bot] opened Oct 30, 2026Bot
1
fix(threading): reset per-pool stream allocation counts
#39351maindevhoon opened Oct 30, 2026
1
[GPU] Fix SimpleGPU custom layers with dynamic outputs loaded from th…
#39347TheSanchezian opened Oct 29, 2026
1
[GPU] Compile identical custom layer kernels once
#39346TheSanchezian opened Oct 29, 2026
1
[NPUW] Map weights from the handle provider's seek position
#39345akodanka opened Oct 29, 2026Contributor 2026.5
6
[CPU] Avoid rebuilding StridedSlice executor when input shapes are unchanged
#39344muntasir-396 opened Oct 29, 2026
3
[CPU] Register GatedDeltaNet in shape inference factory (#39339)
#39343muntasir-396 opened Oct 29, 2026
5
[Transformations] QSAIndexer fusion transformation
#39342mryzhov opened Oct 29, 2026Contributor
2
[HETERO] Add HSM support in Hetero Plugin
#39341praasz opened Oct 29, 2026Contributor 2026.5
1
[GPU] MoE offload: fewer expert reloads in the per-expert prefill fallback
#39338blairducrayoppat opened Oct 29, 2026
1
[NPUW] Validate imported MoE param mapping ports
#39335darktears opened Oct 29, 2026Contributor
4
[Core] hsm: Implement codec for compile options
#39332t-jankowski opened Oct 29, 2026Contributor
25
[NPU] Adding support for the QoS feature
#39330pereanub opened Oct 29, 2026Contributor