lhutton1 opened a new pull request #6430: URL: https://github.com/apache/incubator-tvm/pull/6430
Allows the user to specify a 'custom' label on the operators they wish to use a separate function to determine the layout, rather than relying on purely the operator name. One use-case example of this is were one might require a different layout for depth-wise convolution and other types of convolution. e.g. IHWO for depthwise, else OHWI. See convert_layout.rst for a more detailed explanation. cc @anijain2305 @manupa-arm @mbaret ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
