Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Make type check errors more understandable #136
Recently added type-checking is very useful, but it looks not easy to read. This PR intends to make it more understandable by showing where the error is raised.
Assuming the following model (derived from MNIST example).
After this PR
If base idea looks OK, I'll work on:
referenced this pull request
Nov 6, 2015
recently when we were trying to do inference test for googlenet_v2, similar problem occurred. Would you please help have a look at this issue? @unnonouno
Expect: in_types.shape == in_types.shape
It suggests that "second dimension of first argument must be same as second dimension of second argument in LinearFunciton, but the former is 128 and later is 512". The first is