Search results

[Build] Shared Library QNN not supported on Android, but required to use plug-in style EP

buildbuild issues; typically submitted using templatebuild issues; typically submitted using template

ep:QNNissues related to QNN exeution providerissues related to QNN exeution provider

platform:mobileissues related to ONNX Runtime mobile; typically submitted using templateissues related to ONNX Runtime mobile; typically submitted using template

Status: Open.

#28213 In microsoft/onnxruntime;

· john-dance opened on Apr 23, 2026

Support advanced EP configuration (provider options, config entries) in ONNX backend API

Status: Open.

#28209 In microsoft/onnxruntime;

· titaiwangms opened on Apr 23, 2026

[Build] ORT_API_MANUAL_INIT

buildbuild issues; typically submitted using templatebuild issues; typically submitted using template

platform:webissues related to ONNX Runtime web; typically submitted using templateissues related to ONNX Runtime web; typically submitted using template

Status: Open.

#28207 In microsoft/onnxruntime;

· NazaRik555 opened on Apr 23, 2026

com.microsoft.RotaryEmbedding: returns if position_ids are out of boundary

Status: Open.

#28205 In microsoft/onnxruntime;

· xadupre opened on Apr 23, 2026

add a threshold above which constant folding is disabled

Status: Open.

#28203 In microsoft/onnxruntime;

· xadupre opened on Apr 23, 2026

[Feature Request] Add int4/uint4 support for IBM Power Systems (PowerPC; VSX/MMA)

feature requestrequest for unsupported feature or enhancementrequest for unsupported feature or enhancement

Status: Open.

#28199 In microsoft/onnxruntime;

· mgiessing opened on Apr 23, 2026

[Feature Request] CUDA EP: Support head_dim > 256 in GroupQueryAttention kernel

ep:CUDAissues related to the CUDA execution providerissues related to the CUDA execution provider

Status: Open.

#28196 In microsoft/onnxruntime;

· justinchuby opened on Apr 22, 2026

CUDA EP: Unfused Attention runner produces NaN for fp16 with head_dim > 256

ep:CUDAissues related to the CUDA execution providerissues related to the CUDA execution provider

Status: Open.

#28195 In microsoft/onnxruntime;

· justinchuby opened on Apr 22, 2026

Feature request: GroupQueryAttention support for KV-shared layers (Gemma4-style)

Status: Open.

#28188 In microsoft/onnxruntime;

· justinchuby opened on Apr 22, 2026

[Feature Request] CoreML EP: add QuickGelu support

ep:CoreMLissues related to CoreML execution providerissues related to CoreML execution provider

ep:QNNissues related to QNN exeution providerissues related to QNN exeution provider

model:transformerissues related to a transformer model: BERT, GPT2, Hugging Face, Longformer, T5, etc.issues related to a transformer model: BERT, GPT2, Hugging Face, Longformer, T5, etc.

platform:mobileissues related to ONNX Runtime mobile; typically submitted using templateissues related to ONNX Runtime mobile; typically submitted using template

Status: Open.

#28183 In microsoft/onnxruntime;

· maxwbuckley opened on Apr 22, 2026

[Feature Request] CoreML EP: add HardSigmoid support

ep:CoreMLissues related to CoreML execution providerissues related to CoreML execution provider

platform:mobileissues related to ONNX Runtime mobile; typically submitted using templateissues related to ONNX Runtime mobile; typically submitted using template

Status: Open.

#28181 In microsoft/onnxruntime;

· maxwbuckley opened on Apr 22, 2026

[Performance][CoreML] GatherOpBuilder rejects rank-0 (scalar) indices, forcing CPU fallback for StyleGAN-family models

ep:CoreMLissues related to CoreML execution providerissues related to CoreML execution provider

Status: Open.

#28180 In microsoft/onnxruntime;

· maxwbuckley opened on Apr 22, 2026