All pull requests

Pull requests

• [LFM2] Clear short-conv state at the start of a new sequence

  #23401·alpharomercoma opened Oct 3, 2026·

  ·
  1

  5

• Fix Inspector plot crash on non-finite metrics

  CLA Signedrelease notes: devtools
  #23400·u7k4rs6 opened Oct 3, 2026Contributor·

  ·
  3

• Fix QuantizedKVCache abort / silent corruption on non-float32 activations

  CLA Signed
  #23399·tritsystem opened Oct 3, 2026·

  ·
  2

• MLX: add gather and scatter handlers

  #23398·zsun6 opened Oct 3, 2026·

  ·
  4

• Fix Running on Device link in CUDA backend overview

  CLA Signedrelease notes: none
  #23397·pratikgx opened Oct 3, 2026·

  ·
  3

• Bump datasets from 3.6.0 to 5.0.1 in /examples/openvino/stable_diffusion

  CLA Signeddependenciespython
  #23396·dependabot[bot] opened Oct 3, 2026Bot·

  ·
  2

• [CoreAI] Add root and PTE cache clearing

  CLA Signed
  #23395·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Backend-local PTE inspection

  CLA Signed
  #23394·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Integrate the delegate execute path

  CLA Signed
  #23393·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Coordinate model acquisition

  CLA Signed
  #23392·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Add the Core AI SDK Swift and Objective-C bridge

  CLA Signed
  #23391·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Prepare source bundles

  CLA Signed
  #23390·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Persist raw bookmarks

  CLA Signed
  #23389·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Add atomic cache storage primitives

  CLA Signed
  #23388·metascroy opened Oct 3, 2026Contributor·

  ·
  1

• [CoreAI] Parse the delegate manifest at runtime and run host tests in CI

  CLA Signed
  #23387·metascroy opened Oct 3, 2026Contributor·

  ·
  2

• [RFC] Qualcomm AI Engine Direct - RFCs for Dataset calibration support and LoRA with LLMConfig

  CLA Signed
  #23386·qti-mmadhava opened Oct 3, 2026Contributor·

  ·
  2

• Merge cuda-perf.yml into cuda.yml and add 5% perf regression gate (#23385)

  CLA Signedmeta-exported
  #23385·Gasoonjia opened Oct 3, 2026Contributor·

  ·
  3

• [LLM] Add Terminal-Bench evaluation scripts

  CLA Signed
  #23384·mergennachin opened Oct 3, 2026Contributor·

  ·
  5

• [ET-VK][shaders] Harden UBO vector indexing for Adreno

  CLA Signedmeta-exported
  #23381·SS-JIA opened Oct 3, 2026Contributor·

  ·
  2

• [cadence] Run one Cadence job tree for branch PRs and fork PRs

  CLA Signed
  #23379·aliafzal opened Oct 2, 2026Contributor·

  ·
  3

• Fix codegen

  CLA Signedmeta-exported
  #23378·metascroy opened Oct 2, 2026Contributor·

  ·
  3

• Arm backend: Add static floating-point TopK lowering

  ciflow/trunkCLA Signedmodule: armpartner: armrelease notes: arm
  #23377·YufengShi-dudu opened Oct 2, 2026Collaborator·

  ·
  3

• [executorch][muse] Move off-graph KV to the batching path and test batching in OSS CI

  CLA Signedmeta-exported
  #23375·Gasoonjia opened Oct 2, 2026Contributor·

  ·
  2

• [executorch][muse] Add run_solo_batching

  CLA Signedmeta-exported
  #23374·Gasoonjia opened Oct 2, 2026Contributor·

  ·
  2

• [executorch][muse] Add export_solo_batching.py

  CLA Signedmeta-exported
  #23373·Gasoonjia opened Oct 2, 2026Contributor·

  ·
  2