All pull requests

Pull requests

• Add gemma4 OpenVINO NPU support

  #746·xiaoyu-work opened Sep 23, 2026·

  ·
  7

• Support mixed-width QMoE export

  #744·titaiwangms opened Sep 21, 2026·

  ·
  24

• Separate target build contracts from graph rewrites

  #743·xiaoyu-work opened Sep 21, 2026·

  ·
  4

• Add Qwen2.5 Omni Talker export

  #742·rui-ren opened Sep 18, 2026·

  ·
  11

• [TensorRT] Add static-cache export support and numerical runtime coverage

  #741·rui-ren opened Sep 17, 2026·

  ·
  7

• Add experimental Qwen3.8 packed hybrid export for ORT GenAI Engine

  #738·titaiwangms with CopilotAI opened Sep 16, 2026·

  ·
  1

  14

  

• Bump ruff from 0.16.2 to 0.16.6 in /requirements/lintrunner

  dependenciespython
  #734·dependabot[bot] opened Sep 11, 2026·

  ·
  3

• Add MatterGen crystal diffusion score-core export

  #731·justinchuby opened Sep 4, 2026·

  ·
  22

• Add TimesFM 3 multivariate forecasting support

  #718·justinchuby opened Sep 1, 2026·

  ·
  8

• Add GLM-5.3 Flash multimodal support

  automated
  #713·justinchuby opened Aug 31, 2026·

  ·
  5

• Export DeepSeek V4 ratio-4 compressed sparse attention

  #700·justinchuby opened Aug 29, 2026·

  ·
  2

• Make Qwen4Exp download diagnostics truthful

  #698·justinchuby opened Aug 28, 2026·

  ·
  5

• Enable direct GLM-5.2 GGUF tensor mapping

  #682·justinchuby opened Aug 27, 2026·

  ·
  5

• Tied quantized lm_head fix

  #671·neilmsft opened Aug 27, 2026·

  ·
  4

  

• fix(deepseek-v4): per-layer CSA/HCA compressed-record symbolic dim

  #622·justinchuby opened Aug 25, 2026·

  ·
  3

• Add NVIDIA Nemotron 3.5 Lightning export support

  #487·justinchuby opened Aug 14, 2026·

  ·
  19

• Add WeatherNext-style ONNX conversion demo

  #469·justinchuby with CopilotAI opened Aug 10, 2026·

  ·
  1

  9

  

• Add composable world-model pipelines and Cosmos3 export

  #466·xiaoyu-work opened Aug 8, 2026·

  ·
  31

  

• Fix gemma4 audio-encoder audio_features export rank (#570)

  #444·justinchuby opened Aug 1, 2026·

  ·
  7

• Add streaming multi-talker Parakeet RNN-T ASR model

  #411·themason2011 opened Jul 17, 2026·

  ·
  4

• chore(deps): Update lintrunner-adapters requirement from >=0.14.0 to >=0.14.1 in /requirements/lintrunner

  dependenciespython
  #409·dependabot[bot] opened Jul 17, 2026·

  ·
  2

• Quantize Gemma4 per-layer embedding tables for WebGPU (INT4/INT8 configurable)

  #403·feich-ms opened Jul 14, 2026·

  ·
  32

• Dashboard: collapse per-level status badges into a single highest-level badge

  #347·tadani3 with CopilotAI opened Jun 9, 2026·

  ·
  1

  6

  

• Onboard Qwen2.5-Omni model

  #193·rui-ren opened Apr 23, 2026·

  ·
  27

• Add MaskedLMTask for protein/bio models (ESM-2, ProtBert)

  ai
  #86·justinchuby opened Apr 2, 2026·

  ·
  9