Tong Chen

working on Resize.

The type verifier is generated automatically from .td file. Should we extend the type description to AnyInteger, AnyFloat or just AnyType? I checked this model. It just specified wrong type...

This error is caused by that the name 'conv2d/bias/read' appears in the input list and value_info list . I do not why this error did not appear previously. However, the...

I built on Mac natively with cmake. For the new onnx, I need to specify where my python is: ``` cmake -DPYTHON_EXECUTABLE:FILEPATH=/Library/Frameworks/Python.framework/Versions/3.8/bin/python3 -DPYTHON_LIBRARIES=/Library/Frameworks/Python.framework/Versions/3.8/lib/libpython3.8.dylib .. ``` Not sure whether this is...

The error report showed that two types of error for shape inference: not implemented or unranked tensor. Any way to know which operations the error occurred? Perhaps we should change...

@chenqiny My output already has the operation information in the next line of the error message if emitError is used. Is that the case for your output? The emitError is...

I assume that this approach works only when the transpose is the only user for the input tenor.

@shizy818 Could you attach the model?

@shizy818 For the pytorch model, the error came from the customized op. Its output type is incorrect. May need special handling. For the tensor flow model, the error came from...

@shizy818 @caoimhinuibrian When we use one CustomOp to represent all types of customized ops, one piece of information is missing: type inference for output type. When we construct Ops in...