Search results

Remove prefill calibration

Status: Open.

Task

#17827 In pytorch/executorch;

· abhinaykukkadapu opened on Mar 3, 2026

ET speech/voice models

Status: Open.

#17826 In pytorch/executorch;

· larryliu0820 opened on Mar 3, 2026

add stacktrace in numeric_descrepancy result to map the error back to eager model

Status: Open.

#17825 In pytorch/executorch;

· Gasoonjia opened on Mar 3, 2026

GTC Demo preparation

Status: Open.

#17824 In pytorch/executorch;

· larryliu0820 opened on Mar 3, 2026

Add CI for CortexM Ops (FVP)

Status: Open.

#17822 In pytorch/executorch;

· psiddh opened on Mar 3, 2026

Add support for torch.complex64 on vulkan backend

module: vulkanIssues related to the Vulkan delegate and code under backends/vulkan/Issues related to the Vulkan delegate and code under backends/vulkan/

Status: Open.

#17791 In pytorch/executorch;

· msluszniak opened on Mar 2, 2026

Skip calibrating with generated tokens in the calibration loop.

module: qnnIssues related to Qualcomm's QNN delegate and code under backends/qualcomm/Issues related to Qualcomm's QNN delegate and code under backends/qualcomm/

partner: qualcommFor backend delegation, kernels, demo, etc. from the 3rd-party partner, QualcommFor backend delegation, kernels, demo, etc. from the 3rd-party partner, Qualcomm

Status: Open.

Task

#17785 In pytorch/executorch;

· abhinaykukkadapu opened on Mar 2, 2026

QNN LLM Compilation Pipeline — Profiling & Optimization

module: qnnIssues related to Qualcomm's QNN delegate and code under backends/qualcomm/Issues related to Qualcomm's QNN delegate and code under backends/qualcomm/

module: quantizationIssues related to quantizationIssues related to quantization

partner: qualcommFor backend delegation, kernels, demo, etc. from the 3rd-party partner, QualcommFor backend delegation, kernels, demo, etc. from the 3rd-party partner, Qualcomm

Status: Open.

Task

#17784 In pytorch/executorch;

· abhinaykukkadapu opened on Mar 2, 2026

Qwen3-1.7B to SA8295 fails in dockerized build: num_sharding=1 hits QNN graph memory limit, num_sharding=2 causes host OOM

Status: Open.

#17782 In pytorch/executorch;

· FouadSakr opened on Mar 2, 2026

Provide non-deprecated API for directly accessing input/output tensor memory

module: runtimeIssues related to the core runtime and code under runtime/Issues related to the core runtime and code under runtime/

Status: Open.

Feature

#17781 In pytorch/executorch;

· etrommer opened on Mar 2, 2026

[ARM] GRU fails to fallback to CPU (Non-passthrough operation could not run on NPU)

module: armIssues related to arm backendIssues related to arm backend

partner: armFor backend delegation, kernels, demo, etc. from the 3rd-party partner, ArmFor backend delegation, kernels, demo, etc. from the 3rd-party partner, Arm

Status: Open.

#17753 In pytorch/executorch;

· sarah-aynak opened on Feb 26, 2026

Security

Status: Open.

#17738 In pytorch/executorch;

· lucylq opened on Feb 26, 2026