-
Notifications
You must be signed in to change notification settings - Fork 409
OpenAug 26, 2026
No due date
•Last updated 72% complete
List view
0 of 14 selected 0 issues of 14 selected
Bug Report: Hard-coded CUDA 12 check blocks CUDA 13 support in Torch-TensorRT
bugSomething isn't workingSomething isn't workingStatus: Open.#4438 In pytorch/TensorRT;✨[Feature] Preserve Higher-Level Operators in Torch-TRT Dynamo Paths
feature requestNew feature or requestNew feature or requeststory: Dynamo Frontend & Partitioningtorch.compile, torch.export, FX graph tracing, graph partitioner, graph breaks, and the Dynamo-to-TRtorch.compile, torch.export, FX graph tracing, graph partitioner, graph breaks, and the Dynamo-to-TRStatus: Open.#2729 In pytorch/TensorRT;- Status: Open.#4466 In pytorch/TensorRT;
Fix/converter dtype promotion
component: api [Python]Issues re: Python APIIssues re: Python APIcomponent: conversionIssues re: Conversion stageIssues re: Conversion stagecomponent: convertersIssues re: Specific op convertersIssues re: Specific op converterscomponent: coreIssues re: The core compilerIssues re: The core compilercomponent: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` pathscomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).- Status: Open (in progress).
fix: add CUDA 13 support for TRT-LLM plugins
component: api [Python]Issues re: Python APIIssues re: Python APIcomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(dynamo): keep the complex decomposition inside the TRT op set
component: api [Python]Issues re: Python APIIssues re: Python APIcomponent: coreIssues re: The core compilerIssues re: The core compilercomponent: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` pathscomponent: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passescomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(dynamo): stop unpacking complex inputs before the refit output check
component: api [Python]Issues re: Python APIIssues re: Python APIcomponent: coreIssues re: The core compilerIssues re: The core compilercomponent: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` pathscomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(runtime): give the TensorRT engine class an __eq__
component: coreIssues re: The core compilerIssues re: The core compilercomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(runtime): back a python engine runtime cache with the python handle
component: api [Python]Issues re: Python APIIssues re: Python APIcomponent: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(tests): compare baddbmm lowering against TensorRT in fp32, not TF32
component: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(tests): download the shared torchvision checkpoints before pytest forks
component: testsIssues re: TestsIssues re: TestsStatus: Open (in progress).fix(executorch): share the C++ runtime instead of shipping a second copy
component: api [Python]Issues re: Python APIIssues re: Python APIStatus: Open (in progress).seperate executorch build test in ci
component: api [Python]Issues re: Python APIIssues re: Python APIStatus: Open (in progress).