Dwayne Robinson

/azp run Linux OpenVINO CI Pipeline, orttraining-ortmodule-distributed

/azp run Linux OpenVINO CI Pipeline

> the issue is coming from the graph fusion system I wonder if specifying a lower optimization level like [GraphOptimizationLevel](https://onnxruntime.ai/docs/api/python/api_summary.html#onnxruntime.GraphOptimizationLevel) like `ORT_ENABLE_BASIC` would mitigate the issue until it can be...

> @fdwr is there any more information I can provide that would help with diagnosing/fixing this? Are there any DirectML debug layer messages, if you enable it? https://github.com/microsoft/onnxruntime/issues/13330#issuecomment-1284817983 https://github.com/microsoft/onnxruntime/issues/15255#issuecomment-1487703350

Oooh, so there's a 5D matmul in this model then? ``` D3D12 ERROR: An invalid dimension count of 5 was specified in tensor 'A' which is not between 2 and...

> Edit: I've uploaded the ONNX file to [Google Drive](https://drive.google.com/file/d/1KIDIdNS_QDZfey-9jH8bEOS9JV1hqKrS/view?usp=sharing) - ~~Thanks - downloading now.~~ - Downloaded. - I can repro it locally. 

Reduced to minimal repro, a single operator .onnx file: [minimal-repro.zip](https://github.com/user-attachments/files/16914429/minimal-repro.zip)  Opened bug for DirectML.dll. I'll see what the response is, but in the meantime, we should probably *attempt* to...

> [@fdwr](https://github.com/fdwr) out of interest: is there a way to track the bug that you opened for DirectML? Related to this: you mentioned that ONNXRuntime should attempt to flatten the...

> Thanks for letting me know. Looking forward to when the fix is available. @marovira Alas, it never came out, as the DirectML team was dispersed to the winds in...

Additionally the current [computation here](https://github.com/onnx/onnx/blob/b5e1378a35b24a4316900c69b71214cac7ca5736/onnx/defs/tensor/utils.cc#L137)... ``` int64_t dim_value = static_cast(std::floor(static_cast(input_shape.dim(i).dim_value()) * scales_data[i])); ``` ...has an issue that some output sizes cannot be faithfully represented, as performing the computation in float32...