Jeremie Vandenplas
Jeremie Vandenplas
> Since this approach runs the optimizer on a layer-by-layer basis, optimizer memory such as velocity, rms gradients, etc. are not implemented yet, as they originally assumed whole-network update. Another...
I think it is a good idea. But I have no experience with these tools.
The Fortran-lang project [test-drive](https://github.com/fortran-lang/test-drive/pull/43) also uses such tools.
Thank you @jalvesz . Nice approach.