All pull requests

Pull requests

• fix(conversion): 0-dim tensors, regressions from #4978, attention mask partitioning, float64 ExecuTorch export, and x ** 0.5 as sqrt

  cla signedcomponent: api [Python]component: conversioncomponent: converterscomponent: corecomponent: dynamocomponent: loweringcomponent: tests
  #4995·shoumikhin opened Nov 7, 2026Contributor·

  ·
  1

• test: cover TensorRT-RTX runtime feature usage

  backend: TensorRT-RTXci:trt-apicla signedcomponent: tests
  #4986·tp5uiuc opened Nov 5, 2026Collaborator·4 of 5 tasks·

  ·
  1

  

• fix(executorch): replay CUDA graphs on caller buffers and green contexts, order shared scratch calls with a GPU flag, and check tensor types

  ci: fullcla signedcomponent: api [C++]component: testsdocumentation
  #4983·shoumikhin opened Nov 5, 2026Contributor·

  ·

  · v2.15.0

• fix(executorch): accept ExecuTorch's PyPI CUDA wheel in source builds

  cla signedcomponent: api [Python]component: tests
  #4976·shoumikhin opened Nov 5, 2026Contributor·

  ·

• feat(agentic workflow): add a policy-controlled MCP adapter

  cla signedcomponent: api [Python]component: build systemcomponent: conversioncomponent: corecomponent: dynamocomponent: testsdocumentation
  #4974·zewenli98 opened Nov 5, 2026Collaborator·2/2·7 tasks done·

  ·
  2

• [2.15] Fix TorchAO example failures from the rc3 sanity check

  ci: fullcla signedcomponent: api [Python]component: build systemcomponent: corecomponent: dynamocomponent: loweringcomponent: tests
  #4973·micwill755 opened Nov 4, 2026Collaborator·5 of 7 tasks·

  ·

  · v2.15.0
  1

• Add NVIDIA Python Libs to rpath in setup.py

  cla signedcomponent: build system
  #4970·5had3z opened Nov 3, 2026Collaborator·3 of 7 tasks·

  ·

• fix(partitioning): recompute large input-only values in each resource-split engine

  ci: fullcla signedcomponent: api [Python]component: corecomponent: dynamocomponent: tests
  #4961·cehongwang opened Oct 30, 2026Collaborator·5 of 7 tasks·

  ·

  · v2.15.0
  5

• fix(runtime): allocate TorchScript engines inside libtorchtrt

  cla signedcomponent: api [Python]component: corecomponent: runtimecomponent: tests
  #4960·apbose opened Oct 30, 2026Collaborator·

  ·
  1

• examples(llm): Fixes for FP8 GQA and add Alpamayo FP8 ModelOpt to examples.

  cla signedcomponent: api [Python]component: corecomponent: dynamocomponent: loweringcomponent: tests
  #4959·5had3z opened Oct 30, 2026Collaborator·6 of 7 tasks·

  ·

• fix(kernels) - Avoid TRT Plugin False-Positive Error Logs

  cla signedcomponent: api [Python]component: tests
  #4957·5had3z opened Oct 30, 2026Collaborator·5 of 7 tasks·

  ·

• fix: preserve frozen constant input signatures during export

  cla signedcomponent: api [Python]component: corecomponent: dynamocomponent: tests
  #4952·970334745 opened Oct 30, 2026·5 tasks done·

  ·
  1

  4

• build(deps): bump setuptools from 83.0.0 to 84.1.0 in /toolchains/jp_workspaces

  cla signeddependenciespython
  #4945·dependabot[bot] opened Oct 29, 2026Bot·

  ·

• feat(dynamo): low_memory - compile from a weights file in one block, and release mapped weights by default

  ci: fullcla signedcomponent: api [Python]component: corecomponent: dynamocomponent: runtimecomponent: testsdocumentation
  #4942·cehongwang opened Oct 29, 2026Collaborator·6 of 7 tasks·

  ·

  · v2.15.0
  2

• fix(runtime): keep TRTEngine's layout independent of build macros

  ci: nightlycla signedcomponent: corecomponent: runtimecomponent: tests
  #4940·ConnorBaker opened Oct 29, 2026·4 of 7 tasks·

  ·
  7

  

• feat: support aten.masked_select by decomposing to boolean indexing

  cla signedcomponent: api [Python]component: corecomponent: dynamocomponent: loweringcomponent: tests
  #4935·Baiiiiiiiiii opened Oct 27, 2026·5 of 7 tasks·

  ·
  1

  4

• test(runtime): build the compile-memory test engines in fp32

  cla signedcomponent: tests
  #4932·narendasan opened Oct 26, 2026Collaborator·

  ·

• feat : enable TRT API model tests for TensorRT-RTX

  backend: TensorRT-RTXci:trt-apicla signedcomponent: tests
  #4926·tp5uiuc opened Oct 23, 2026Collaborator·6 tasks done·

  ·
  1

  

• fix : skip FP8 MHA accuracy checks on TRT-RTX 1.7 and earlier

  ci: fullcla signedcomponent: tests
  #4925·tp5uiuc opened Oct 23, 2026Collaborator·2 of 7 tasks·

  ·
  1

  

• fix: validate 3D convolutions on Turing with TensorRT-RTX

  backend: TensorRT-RTXcla signedcomponent: api [Python]component: conversioncomponent: converterscomponent: corecomponent: dynamocomponent: testsdocumentation
  #4924·tp5uiuc opened Oct 23, 2026Collaborator·6 tasks done·

  ·

  · v2.16.0
  

• fix: correct TensorRT-RTX converter tests on Turing and A30

  ci: fullcla signedcomponent: tests
  #4923·tp5uiuc opened Oct 23, 2026Collaborator·5 tasks done·

  ·

  · v2.16.0
  2

  

• feat(agentic workflow): add structured reporting, verification, and benchmarking workflows for agents

  cla signedcomponent: api [Python]component: build systemcomponent: conversioncomponent: corecomponent: dynamocomponent: testsdocumentation
  #4920·zewenli98 opened Oct 22, 2026Collaborator·1/2·7 tasks done·

  ·
  

• feat: agentic region kernel generation

  cla signedcomponent: api [Python]component: build systemcomponent: conversioncomponent: corecomponent: dynamocomponent: loweringcomponent: runtimecomponent: testsdocumentation
  #4914·bowang007 opened Oct 22, 2026Collaborator·1 of 7 tasks·

  ·
  9

• fix(runtime): create the warm-up inputs before unflattening them under whole-graph CUDA graphs

  cla signedcomponent: api [Python]component: corecomponent: dynamocomponent: runtimecomponent: tests
  #4907·narendasan opened Oct 21, 2026Collaborator·4/4·

  ·

• fix(runtime): keep the output structure under whole-graph CUDA graphs

  cla signedcomponent: api [Python]component: corecomponent: dynamocomponent: runtimecomponent: tests
  #4906·narendasan opened Oct 21, 2026Collaborator·3/4·

  ·