Search results

Feature request: quantized (int8/fp8) KV cache in GroupQueryAttention with past/present shared buffer + CUDA graph capture

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

quantizationissues related to quantizationissues related to quantization

Status: Open.

#29783 In Microsoft/onnxruntime;

· justinchuby opened on Jul 19, 2026

IoBinding + CUDA graph capture: bound CPU inputs require ClearBoundInputs + full re-bind each replay to be observed

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

Status: Open.

#29782 In Microsoft/onnxruntime;

· justinchuby opened on Jul 18, 2026

[MIGraphX EP] Hard RuntimeException (no CPU fallback) when MIGraphX fails to compile an assigned subgraph

ep:MIGraphXissues related to AMD MI GraphX execution providerissues related to AMD MI GraphX execution provider

Status: Open.

#29778 In Microsoft/onnxruntime;

· AIwork4me opened on Jul 18, 2026

[Tracking] Constrained-environment memory and layer-offload roadmap

memory

Status: Open.

#29775 In Microsoft/onnxruntime;

· titaiwangms opened on Jul 18, 2026

[Documentation] CUDA EP documentation doesn't cover CUDA 13

documentationimprovements or additions to documentation; typically submitted using templateimprovements or additions to documentation; typically submitted using template

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

Status: Open.

#29767 In Microsoft/onnxruntime;

· stmccartneyncc opened on Jul 17, 2026

onnxruntime-node 1.14.0 SIGTRAP crash in BFCArena during model load — reproducible in Electron only, not plain Node.js

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

Status: Open.

#29763 In Microsoft/onnxruntime;

· justanotherkevin opened on Jul 17, 2026

symbolic_shape_infer.py crashes on a Range node with non-scalar inputs when fixing dynamic shapes on a dynamo-exported LoRA-adapter graph (--fixed_param_dict)

Status: Open.

#29762 In Microsoft/onnxruntime;

· chkrishna2001 opened on Jul 17, 2026

windows_x64_release_vitisai became very slow starting 7.9

Status: Open.

#29760 In Microsoft/onnxruntime;

· xadupre opened on Jul 17, 2026

Windows build became 2 times slower around 7/9

platform:windowsissues related to the Windows platformissues related to the Windows platform

Status: Open.

#29758 In Microsoft/onnxruntime;

· xadupre opened on Jul 17, 2026

[Feature Request] Compile API: allow TensorRT-RTX compile-only sessions without a live/matching GPU

ep:TensorRTissues related to TensorRT execution providerissues related to TensorRT execution provider

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

Status: Open.

#29757 In Microsoft/onnxruntime;

· matpoudret opened on Jul 17, 2026

Microsoft WebNN demo applications cannot successfully run in ORT-WebGPU backend with CPU fallback disabled

ep:WebGPUort-web webgpu providerort-web webgpu provider

ep:WebNNWebNN execution providerWebNN execution provider

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

Status: Open.

#29756 In Microsoft/onnxruntime;

· qwu16 opened on Jul 17, 2026

CI lint checks install an unused rust toolchain (and it's flaky)

Status: Open.

#29750 In Microsoft/onnxruntime;

· mcollinswisc opened on Jul 17, 2026