t-vi commented on pull request #5779: URL: https://github.com/apache/incubator-tvm/pull/5779#issuecomment-642923575
@masahi So what I did here in addition to what is seen was to locally assert that no `float` is passed to const/var creation as dtype. This is related to the forum discussion here: https://discuss.tvm.ai/t/discuss-the-meaning-of-float-in-relay/6949/ The next step will be to apply the data type of input tensors on scalar inputs, but I'll keep that for a separate PR. In addition to my primary use-case - allowing doubles to facilitate testing of conversions - this likely helps with mixed precision etc. ---------------------------------------------------------------- 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]
