Search results

Feature request: Implement GroupNormalization-21

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

core runtimeissues related to core runtimeissues related to core runtime

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

Status: Open.

#24538 In microsoft/onnxruntime;

· justinchuby opened on Apr 24, 2025

Squeeze node fails when axes is ""

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

Status: Open.

#21661 In microsoft/onnxruntime;

· justinchuby opened on Aug 7, 2024

Pad behavior is inconsistent with ONNX ReferenceEvaluator and documentation

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

core runtimeissues related to core runtimeissues related to core runtime

Status: Open.

#21037 In microsoft/onnxruntime;

· vincentme opened on May 24, 2024

LayerNormalization doesnt' work as expected on Mac

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

core runtimeissues related to core runtimeissues related to core runtime

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

Status: Open.

#20676 In microsoft/onnxruntime;

· fatcat-z opened on May 14, 2024

[Feature request] Support bfloat16/float8 inputs in <code>session.run()</code>

apiissues related to all other APIs: C, C++, Python, etc.issues related to all other APIs: C, C++, Python, etc.

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#20578 In microsoft/onnxruntime;

· justinchuby opened on May 6, 2024

[dynamo_export] MLFloat16 data type is not supported with ScatterElements opset 18 when reduction is 'max'.

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#18864 In microsoft/onnxruntime;

· BowenBao opened on Dec 18, 2023

[dynamo_export] starts_.size() == ends_.size() + 1 was false. No matching 'start' entry.

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#18863 In microsoft/onnxruntime;

· BowenBao opened on Dec 18, 2023

[tracking] Improve Sequence operator handling

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

core runtimeissues related to core runtimeissues related to core runtime

Status: Open.

#18355 In microsoft/onnxruntime;

· justinchuby opened on Nov 8, 2023

op.SequenceEmpty(dtype=xxx) cannot be set to float16.

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#16846 In microsoft/onnxruntime;

· xiaowuhu opened on Jul 25, 2023

ORT aborts on ConcatFromSequence with empty sequence inputs

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

core runtimeissues related to core runtimeissues related to core runtime

Status: Open.

#16619 In microsoft/onnxruntime;

· justinchuby opened on Jul 7, 2023

ONNXRuntimeError: Training mode does not support BN opset 14 (or higher) yet.

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#16867 In microsoft/onnxruntime;

· BowenBao opened on Jun 29, 2023

Setting <code>CUBLAS_WORKSPACE_CONFIG=":4096:8"</code> leads to <code>CUBLAS_STATUS_ALLOC_FAILED</code>

converter:dynamoissues related supporting the PyTorch Dynamo exporterissues related supporting the PyTorch Dynamo exporter

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

Status: Open.

#16512 In microsoft/onnxruntime;

· BowenBao opened on Jun 28, 2023