All pull requests

Pull requests

• [WebGPU] Move full provider test coverage to shared plugin CI

  #33269·Yang Gu (gyagp) opened Oct 11, 2026Contributor·

  ·
  4

• Stop 1DS telemetry before SDK static destruction

  #33267·赤坂理子 (akasakariko) opened Oct 11, 2026·

  ·
  1

  3

• [MLAS] Skip FP16 unit tests on CPUs without FP16 arithmetic

  #33260·Yair Ben Avraham (yairbenavraham) opened Oct 11, 2026Contributor·

  ·
  4

• Fix builds with newer installed ONNX versions

  #33259·Rui Chen (chenrui333) opened Oct 11, 2026·

  ·
  2

• [MLAS] Verify AMX tile-state preservation before QGEMM dispatch

  #33258·tianleiwu with CopilotAI opened Oct 11, 2026·

  ·
  1

  1

  

• [WebGPU] Support fixed subgroup size in im2col_matmul

  #33249·Jiawei Shao (Jiawei-Shao) opened Oct 10, 2026Contributor·

  ·
  3

• [CUDA] Add opt-in SM121 L2 prefetch for quantized decode

  #33248·kunal-vaishnavi opened Oct 10, 2026Contributor·

  ·

• webgpu: Fix Python factory selection for plugin EP hosts

  #33241·Jiajia Qin (qjia7) opened Oct 10, 2026Contributor·

  ·
  13

• [CUDA] Plumb GQA workspace envelope into Level-2 declaration

  #33240·Chi Lo (chilo-ms) opened Oct 9, 2026Contributor·

  ·

• [WIP] Fix float16 weights quantization to avoid zeroing out tensors

  #33239·tianleiwu with CopilotAI opened Oct 9, 2026·

  ·
  1

  1

  

• Add opt-in platform-specific onnxruntime-node packages

  #33237·Andrew Ovens (avens19) opened Oct 9, 2026·

  ·
  4

• Reject overflowing Concat axis dimensions

  #33235·Xavier Dupré (xadupre) opened Oct 9, 2026Member·

  ·
  3

• Remove stale Intel MKL third-party notice

  #33233·Xavier Dupré (xadupre) opened Oct 9, 2026Member·

  ·
  1

  1

• [CUDA][WebGPU] Add PackedSparseAttentionIndexerMerge for MTP index sharing

  #33224·kunal-vaishnavi opened Oct 9, 2026Contributor·

  ·
  1

• [Objective-C] Add encrypted compiled EPContext inference coverage

  #33223·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  9

• [React Native] Add opt-in encrypted compiled EPContext device workflow

  #33222·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  7

• Test Python encrypted compiled-model inference

  #33218·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  11

• Test Node.js encrypted compiled-model inference

  #33219·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  6

• Test Java encrypted compiled-model inference

  #33217·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  12

• Test Rust encrypted compiled-model inference

  #33220·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  6

• Test C# encrypted compiled-model inference

  #33221·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  6

• Add executable EPContext fixture for encrypted binding tests

  #33216·Gopalakrishnan Nallasamy (GopalakrishnanN) opened Oct 8, 2026Contributor·

  ·
  10

• [WebGPU] Add GatedDeltaNetStateReplay contrib operator

  #33213·aciddelgado opened Oct 8, 2026Contributor·

  ·
  1

• [WebGPU EP] fuse layer norm projections into gated delta net

  #33211·Prathik Rao (prathikr) opened Oct 8, 2026Contributor·

  ·
  6

• Update libprotobuf-mutator to handle protobuf api changes

  #33210·eserscor opened Oct 8, 2026Contributor·

  ·
  3