Pull Requests microsoft/onnxruntime

Replace -infinity by lowest for masked values

#26055 opened 2025-09-16 13:40 by xadupre

Fix: Replace macOS-13 with macOS-15 in workflows

#26045 opened 2025-09-15 21:18 by snnn

Make local mirror of cmake dependencies configurable

#26042 opened 2025-09-15 10:16 by qti-hungjuiw

Web/WASM: Copy back results for pre‑allocated CPU outputs (fixes #25786)

#26037 opened 2025-09-15 02:13 by Jonahcb

feat(tensorrt): add per-session use_tf32 provider option (#26029)

#26035 opened 2025-09-14 14:05 by Jonahcb

fix(layernorm): allow normalized_size==1 instead of error (#25896)

#26034 opened 2025-09-13 17:50 by Jonahcb

Register SimplifiedLayerNormalization to MS domain

#26033 opened 2025-09-12 23:14 by justinchuby

Debug failing cuda profiler test

#26032 opened 2025-09-12 22:39 by adrianlizarraga

[ARM CPU] SVE support for Sgemm kernel

#26027 opened 2025-09-12 15:07 by akote123

Support plugin EPs with original SessionOptionsAppendExecutionProvider API

#26019 opened 2025-09-11 23:18 by adrianlizarraga

Fix #25890: update value_info when Gemm(transB=1) -> MatMul to avoid …

#26018 opened 2025-09-11 17:02 by Jonahcb

Update PyTorch to version 2.8.0

#26006 opened 2025-09-09 23:57 by snnn

Bump vite from 6.3.5 to 6.3.6 in /js/web/test/e2e/exports/testcases/vite-default platform:web dependencies javascript

#26000 opened 2025-09-09 21:41 by dependabot[bot]

Adding documentation for device tensor creation using CreateMemoryInfo_V2 API in an EP agnostic way

#25995 opened 2025-09-09 11:18 by praneshgo

Update vcpkg to version 2025.08.27

#25993 opened 2025-09-09 06:45 by snnn

[QNN-EP] Enable Optrace from QNN into QNN EP ep:QNN

#25987 opened 2025-09-08 22:06 by quic-calvnguy

Add skip_pip_install to build.py

#25982 opened 2025-09-08 10:39 by qti-hungjuiw

Bump reviewdog/action-shellcheck from 1.31.0 to 1.32.0 dependencies github_actions

#25976 opened 2025-09-08 01:04 by dependabot[bot]

System.Numeric.Tensors breaking change fix

#25972 opened 2025-09-06 21:44 by tjwald

Clean up CUDA contrib ops folder

#25969 opened 2025-09-05 21:46 by hariharans29

[TRT RTX EP] Rename TRT RTX libs to not collide with TRT

#25964 opened 2025-09-05 15:29 by gedoensmax

[webgpu] Split large inputs into smaller buffers to bypass maxStorageBufferBindingSize limit

#25962 opened 2025-09-05 08:56 by xiaofeihan1

[webgpu] Move copy_kv_cache to template

#25961 opened 2025-09-05 06:16 by qjia7

Add onnxruntime_plugin_ep_onnx_test for plugin EPs to run onnx test

#25942 opened 2025-09-04 00:25 by chilo-ms

[WebGPU] allow async shader compilation ep:WebGPU

#25941 opened 2025-09-03 22:56 by fs-eire

[webgpu] Optimize flash decoding by merging QKT and SplitVx shader

#25929 opened 2025-09-03 03:07 by xiaofeihan1

Fix Local Attention off by 1 bug

#25927 opened 2025-09-02 22:26 by aciddelgado

Update AppendExecutionProvider docs

#25924 opened 2025-09-02 20:02 by edgchen1

[Web] Magic import comment missing for Vite (/* @vite-ignore */)

#25919 opened 2025-09-02 11:50 by mcost45

[TRT RTX EP] Remove unused legacy code from TRT EP that is no longer supported for TRT RTX ep:NvRTX

#25916 opened 2025-09-02 09:23 by gedoensmax