Ashwini Khade

Results 71 comments of Ashwini Khade

@lunasdejavu : What are you trying to achieve by converting the onnx model back to pytorch? If this onnx model? If this onnx model is the final model you want...

@InternetExplorer9 : This is a known issue in ONNX Optimizers. There is discussion going on to decide the fate of optimizers. At the very least the call to optimize should...

dilation attribute was added for MaxPool from opset version 10. Are you sure you tried converting to onnx opset 10 or above? FYI @spandantiwari

Please open this issue in the tf2onnx converter repo

Is this related to a tutorial in this repo? If so please add more details. If this is related to general model conversion then please open this issue in tf2onnx...

Will investigate this and report back on my findings

@GuoGuiRong Are you comparing ORT with Triton-ORT? Can you add more details regarding 1. ORT and Triton-ORT configs used durnig testing 2. What perf diff are you seeing

As the error message suggests : MakeCooSparse is not a registered function or operator meaning ORT does not have any operator "MakeCooSparse" registered. Looks like this op is still in...

@supercharleszhu : Can you share a simple model without custom ops like MakeCooSparse and simple test data for same so that we can repro this on our end. Thanks!