Issues

• scan/map: a view of a captured buffer or of the carry inside the body reads unwritten memory

  pytorch/executorch#25075 · Filter by author AxelNoun opened on Nov 14, 2026

• [XNNPACK] upsample_bilinear2d ignores scale_factors when input size × scale is not an integer

  pytorch/executorch#25074 · Filter by author AxelNoun opened on Nov 14, 2026

• [XNNPACK] Statically quantized Linear used twice: the second call gets the first call's quantized bias

  pytorch/executorch#25073 · Filter by author AxelNoun opened on Nov 14, 2026

• [XNNPACK] softmax(dim=-1) on a 4-D tensor after conv/pool is computed over channels

  pytorch/executorch#25072 · Filter by author AxelNoun opened on Nov 14, 2026

• Portable amax/amin abort the process on an input with zero elements

  pytorch/executorch#25071 · Filter by author AxelNoun opened on Nov 14, 2026

• executorch_prim::mod returns a negative result where Python's % is positive (wrong dynamic pad sizes)

  pytorch/executorch#25070 · Filter by author AxelNoun opened on Nov 14, 2026

• CUDA backend: graph replay sync, one graph shape per method, and multi-method buffer errors (found running Alpamayo2-Edge)

  module: cuda
  pytorch/executorch#25067 · Filter by author shoumikhin opened on Nov 14, 2026
  

• llm_server: SessionRuntime.aclose_worker() can spin forever on Python 3.12+

  pytorch/executorch#25051 · Filter by author mergennachin opened on Nov 13, 2026

• Praxis-1 (27B) export to ExecuTorch with TensorRT: how to reproduce from source

  pytorch/executorch#25019 · Filter by author shoumikhin opened on Nov 13, 2026

• Limit per-session prompt-end retention in the shared prefix cache

  pytorch/executorch#24914 · Filter by author mergennachin opened on Nov 11, 2026
  1

• [ET-VK] BUFFER softmax and layer_norm abort on oversized tiled dispatch

  pytorch/executorch#24906 · Filter by author flyingdoog opened on Nov 11, 2026

• CUDA backend: aten._grouped_mm cannot be lowered on A100 (sm_80), so MoE models do not export there

  module: cuda
  pytorch/executorch#24902 · Filter by author shoumikhin opened on Nov 11, 2026
  

• CUDA backend: the default export spends more time on host copies than on the model (MoE layer: 2.3 ms of kernels, 4 ms of copies)

  module: cuda
  pytorch/executorch#24901 · Filter by author shoumikhin opened on Nov 11, 2026
  

• CUDA backend: a program crashes with an illegal memory access when CUDA device 1 is current

  bugmodule: cuda
  pytorch/executorch#24887 · Filter by author shoumikhin opened on Nov 11, 2026

• [QNN] HTP fp16 log_softmax returns shifted values when the softmax axis length is 2^n + 1 (65 … 2049)

  pytorch/executorch#24886 · Filter by author ujma1234 opened on Nov 11, 2026

• samsung device farm stopped assigning devices, test-samsung-models-linux times out on every trunk run

  module: cimodule: samsungpartner: samsung
  pytorch/executorch#24869 · Filter by author shoumikhin opened on Nov 10, 2026
  4

• [RFC] Run a .pte from Python without PyTorch: drop the PyTorch requirement and load CUDA on demand

  pytorch/executorch#24820 · Filter by author shoumikhin opened on Nov 10, 2026
  2

• [RFC] ExecuTorch for ROS 2: a rosdep package, an inference node and a ros2_control controller

  pytorch/executorch#24819 · Filter by author shoumikhin opened on Nov 10, 2026
  1

  

• [RFC] Ship the ExecuTorch C++ SDK as a standalone archive with each release, starting with 1.6.0

  pytorch/executorch#24816 · Filter by author shoumikhin opened on Nov 9, 2026

• Slow small-batch quantized inference on M1/M2 GPUs

  pytorch/executorch#24812 · Filter by author mergennachin opened on Nov 9, 2026

• Speculative decoding loses speed to repeated per-round work

  pytorch/executorch#24811 · Filter by author mergennachin opened on Nov 9, 2026

• CUDA backend: no fallback kernels for nonzero and masked_select

  pytorch/executorch#24767 · Filter by author shoumikhin opened on Nov 7, 2026
  1

  

• XNNPACK packs a reused weight once per use, so memory grows with every reuse

  pytorch/executorch#24766 · Filter by author shoumikhin opened on Nov 7, 2026

• No safe way to set the CPU thread count from Python: _unsafe_reset_threadpool after load_method crashes or hangs XNNPACK

  bug
  pytorch/executorch#24765 · Filter by author shoumikhin opened on Nov 7, 2026

• torch.masked_select fails to_edge ("not in Core ATen opset"), although the portable kernel exists

  bug
  pytorch/executorch#24764 · Filter by author shoumikhin opened on Nov 7, 2026
  1