masahi commented on a change in pull request #9475:
URL: https://github.com/apache/tvm/pull/9475#discussion_r749626598
##########
File path: src/relay/op/image/resize.cc
##########
@@ -215,6 +225,8 @@ bool Resize3DRel(const Array<Type>& types, int num_inputs,
const Attrs& attrs,
const Resize3DAttrs* param = attrs.as<Resize3DAttrs>();
ICHECK(param != nullptr);
+ ICHECK(param->size.size() == 3);
+ ICHECK(param->roi.size() == 10);
Review comment:
Related to above, I wonder if it is safe hard code NCH/NCHW/NCDHW layout
in the ONNX `Resize` op converter, when the spec doesn't care about layout...
https://github.com/apache/tvm/blob/4c1b66f9379fed29988cefd7334f0cd4918c8e6a/python/tvm/relay/frontend/onnx.py#L2612-L2619.
I remember the folklore of ONNX being tied to NCHW layout, but I don't find
the official statement of that convention in the spec.
Is this code supposed to work on models exported from TF?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
