pytorch/TensorRT

View on GitHub

PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT

Stars: 3k
Forks: 410
Open beginner issues: 0
Indexed issues: 231
Avg merge: 3d 18h
Merged PRs (30d): 78

Dominant language: Python
License: BSD-3-Clause
Last GitHub push: Sep 17, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

231 issues indexed so far

Host-resident TensorRT shape-tensor inputs incur a device round trip and a per-call stream sync Open

Difficulty 4/5 3-5 days Newbie friendliness 68/100

pytorch/TensorRT#4717 · Opened Sep 15, 2026
🐛 [Bug] test_base_fp4_static_shapes still borderline-flaky with cosine_similarity Open

bug

pytorch/TensorRT#4697 · 1 comment · 1 assignee · Opened Sep 9, 2026
🐛 [Bug] test_fp8_mha_fused_kernel fails on H100 (root cause undiagnosed) Open

bug

pytorch/TensorRT#4696 · 2 comments · 1 assignee · Opened Sep 9, 2026
🐛 [Bug] `lift_mutated_buffers` crashes on any model that returns a mutated buffer Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 68/100

pytorch/TensorRT#4694 · 1 comment · Opened Sep 9, 2026
✨[Feature] a global switch for engine execution profiling, reachable from an AOTInductor package Open

Difficulty 4/5 3-5 days Newbie friendliness 52/100

pytorch/TensorRT#4679 · 2 comments · Opened Sep 4, 2026
🐛 [Bug] execute_engine input checks name neither the engine nor the binding that failed Open

Difficulty 3/5 1-2 days Newbie friendliness 70/100

pytorch/TensorRT#4678 · 3 comments · Opened Sep 4, 2026
🐛 [Bug] the execute_engine meta kernel reports rank 0 for a sym_size output the engine returns as rank 1 Open

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pytorch/TensorRT#4677 · 2 comments · Opened Sep 4, 2026
🐛 [Bug] inline_torch_modules matches submodule placeholders to parent nodes by name, mis-wiring the graph Open

Difficulty 3/5 1-2 days Newbie friendliness 76/100

pytorch/TensorRT#4612 · Opened Aug 27, 2026
🐛 [Bug] extract_var_range_info calls int() on unbounded range bounds (no lower guard, upper guards only int_oo) Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 64/100

pytorch/TensorRT#4611 · Opened Aug 27, 2026
🐛 [Bug] optimization-profile opt extent is computed as min + max/2, exceeds max value in certain cases Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 55/100

pytorch/TensorRT#4610 · Opened Aug 27, 2026
🐛 [Bug] rank-8+ tensors cross the partition boundary and fail in add_input, not at their producer Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 58/100

pytorch/TensorRT#4608 · Opened Aug 27, 2026
🐛 [Bug] slice does not clamp a finite stop on a dynamic axis, baking the literal stop into the engine Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 68/100

pytorch/TensorRT#4607 · Opened Aug 27, 2026
🐛 [Bug] mul emits ElementWiseOperation.PROD for bool × bool, which TensorRT rejects (AND is supported) Open

bug

pytorch/TensorRT#4605 · 1 assignee · Opened Aug 27, 2026
🐛 [Bug] sort_validator reads an optional argument positionally and raises IndexError Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 68/100

pytorch/TensorRT#4603 · Opened Aug 27, 2026
[CI] torch-compile-models-standard: test_bert_base_uncased failed Open

bug

pytorch/TensorRT#4496 · 1 assignee · Opened Aug 16, 2026
🐛 [Bug] Error with Cask conv when using Hardware Compatible flag Open

bug

pytorch/TensorRT#4481 · 2 comments · 1 assignee · Opened Aug 12, 2026
Dynamo constant folding can leak mutable state through graph outputs Open

pytorch/TensorRT#4466 · 2 comments · 1 assignee · Opened Aug 7, 2026
Bug Report: Hard-coded CUDA 12 check blocks CUDA 13 support in Torch-TensorRT Open

bug

pytorch/TensorRT#4438 · 2 comments · 1 assignee · Opened Jul 28, 2026
Native DistCollective fails at build time when fused into a Myelin ForeignNode (large multi-device graph) Open

pytorch/TensorRT#4381 · 1 comment · 1 assignee · Opened Jul 4, 2026
Backport workflow-hardening fix (artipacked) to 5 release branches Open

Difficulty 3/5 1-2 days Newbie friendliness 72/100

pytorch/TensorRT#4366 · Opened Jun 26, 2026
Support TensorRT weight streaming in the ExecuTorch delegate so large models can run Open

Difficulty 4/5 3-5 days Newbie friendliness 58/100

pytorch/TensorRT#4334 · Opened Jun 10, 2026
🐛 [Bug] Python only pipeline failed on weight streaming test Open

bug

Difficulty 3/5 1-2 days Newbie friendliness 55/100

pytorch/TensorRT#4333 · 2 comments · Opened Jun 10, 2026
🐛 [Bug] Fake Tensor Updater Doesn't Exist on Windows Platform Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 45/100

pytorch/TensorRT#4311 · Opened May 29, 2026
✨[Feature] Support in-place custom plugins (and multiple outputs) Open

feature request

pytorch/TensorRT#4240 · 7 comments · 1 assignee · Opened May 7, 2026
🐛 [Bug] Torch-TensorRT Slower than Onnx TensorRT in Alpamayo-10B Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pytorch/TensorRT#4239 · 1 comment · Opened May 6, 2026
🐛 [Bug] Engine caching and loading from cache Open

bug

pytorch/TensorRT#4226 · 5 comments · 1 assignee · Opened Apr 30, 2026
✨[Feature] Multi-Framework Runner Support in tools Open

feature request

pytorch/TensorRT#4220 · 3 comments · 1 assignee · Opened Apr 28, 2026
convert_exported_program_to_serialized_trt_engine fails with AssertionError on TorchExportableModuleWithStaticCache exports Open

pytorch/TensorRT#4162 · 3 comments · 2 assignees · Opened Apr 2, 2026
[Bug] 17 dynamo test failures across all GPUs on CUDA 13.x — SDP use_fp32_acc, BERT accuracy, AutomaticPlugin, refit Open

bug

Difficulty 5/5 Over a week Newbie friendliness 30/100

pytorch/TensorRT#4154 · 1 comment · Opened Mar 30, 2026
[Bug] MutableTorchTRTModule refit flag stuck at NEEDS_REFIT, never transitions to LIVE on B100/H100 with CUDA 13.x Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 25/100

pytorch/TensorRT#4153 · Opened Mar 30, 2026
[Bug] C++ API and core tests fail with "assert 3 == 0" across all GPUs and CUDA 13.x versions Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pytorch/TensorRT#4152 · 1 comment · Opened Mar 30, 2026
[Bug] TestTorchTensorRTModule::test_get_layer_info AssertionError on H100 with CUDA 13.x Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pytorch/TensorRT#4151 · Opened Mar 30, 2026
🐛 [Bug] add_attention fails with dynamic head dimension when exporting GQA models via torch.export Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pytorch/TensorRT#4145 · 1 comment · Opened Mar 24, 2026
[Bug] All-zero logits from convert_exported_program_to_serialized_trt_engine with use_explicit_typing=False Open

Story: Runtime & Memory & Serialization

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pytorch/TensorRT#4141 · 1 comment · Opened Mar 19, 2026
🐛 [Bug] Encountered bug when using Torch-TensorRT Open

bug story: Dynamo Frontend & Partitioning

Difficulty 4/5 3-5 days Newbie friendliness 45/100

pytorch/TensorRT#4137 · 1 comment · Opened Mar 18, 2026
❓ [Question] Excessive GPU memory usage during constant folding pass causes OOM (Torch-TensorRT + Qwen-Image) Open

question story: Dynamic Shapes & Symbolic Tracing

pytorch/TensorRT#4128 · 2 comments · 1 assignee · Opened Mar 15, 2026
🐛 [Bug] RuntimeError: Failed to extract symbolic shape expressions from source FX graph partition Open

bug story: Dynamic Shapes & Symbolic Tracing

Difficulty 4/5 3-5 days Newbie friendliness 35/100

pytorch/TensorRT#4105 · 2 comments · Opened Mar 2, 2026
🐛 [Bug] opt-shape cannot be set same as min-shape when torch_tensorrt.compile Open

Blocked [PyTorch] bug story: Dynamic Shapes & Symbolic Tracing

Difficulty 3/5 1-2 days Newbie friendliness 45/100

pytorch/TensorRT#4103 · 1 comment · Opened Feb 28, 2026
✨[Feature] Debugging Partitioner Open

feature request story: Dynamo Frontend & Partitioning

pytorch/TensorRT#4100 · 1 comment · 1 assignee · Opened Feb 24, 2026
🐛 [Bug] Incorrect output when using grid-sample operator with use_explicit_typing since tensorrt 10.14.1.48 Open

bug story: Operator Coverage & Converters

Difficulty 4/5 3-5 days Newbie friendliness 35/100

pytorch/TensorRT#4074 · 18 comments · 1 reaction · Opened Feb 7, 2026
qwen3 embedding reranker support? Open

question story: LLM & Generative AI

Difficulty 5/5 Over a week Newbie friendliness 25/100

pytorch/TensorRT#4073 · 1 comment · Opened Feb 6, 2026
🐛 [Bug] run llm exmples are broken in thor platform Open

bug story: LLM & Generative AI

pytorch/TensorRT#4069 · 1 comment · 1 assignee · Opened Feb 4, 2026
🐛 [Bug] flux Examples are broken in thor platform Open

bug story: LLM & Generative AI

pytorch/TensorRT#4068 · 3 comments · 1 assignee · Opened Feb 4, 2026
✨[Feature] Route FP16 Layer Norm properly Open

feature request story: Quantization & Precision

pytorch/TensorRT#4061 · 1 assignee · Opened Jan 30, 2026
↔ [Converter] Add support for torch.ops.aten._local_scalar_dense.default Open

component: converters feature request story: Operator Coverage & Converters

Difficulty 4/5 3-5 days Newbie friendliness 30/100

pytorch/TensorRT#4051 · Opened Jan 27, 2026
📖 [Story] Issue tracker for supporting Alpamayo Open

Story story: LLM & Generative AI

Difficulty 5/5 Over a week Newbie friendliness 10/100

pytorch/TensorRT#4049 · Opened Jan 27, 2026
🐛 [Bug] L1 resource partitioner test error on DLFW Open

bug story: CI/CD & Testing Infrastructure

pytorch/TensorRT#4040 · 1 assignee · Opened Jan 26, 2026
✨[Feature] Improve the backtraces for operators that will fallback Open

feature request story: Dynamo Frontend & Partitioning

pytorch/TensorRT#4022 · 1 assignee · Opened Jan 20, 2026
[aten::embedding] Do we still need this check? Open

story: Operator Coverage & Converters

pytorch/TensorRT#4008 · 1 comment · 1 assignee · Opened Jan 13, 2026
🐛 [Bug] resource split testcase failed in windows Open

bug Story: Build & Install & Packaging

pytorch/TensorRT#3999 · 2 comments · 1 assignee · Opened Jan 7, 2026

pytorch/TensorRT

Get new issues in your inbox