Dingquan

Hi, I'm running into the same issue and found the suggestion by @kazagistar to be very useful. Thanks. One minor issue of that snippet is it will always run whenever...

Thanks. It works. But I think this issue should be remained open until it comforms to numpy.

I think directly returning `self.model.loss_batch(batch)` should be fine. Something looks like: ``` def loss_func(*params): return self.model.loss_batch(batch) ``` User can define whatever they like as the implementation of `model.loss_batch(batch)`, for example,...

On Thu, Mar 16, 2017 at 7:49 PM, Tianjun Xiao wrote: > @wddabc Thanks a lot for the suggestions. We > totally agree on the thought that flags like "isTrain"...

I think there is still flaw of the dry-run idea. I forgot the user might want to build graph themselves and shouldn't implement this. Well, this looks like a hard...

I agree, as I said in the last email. Maybe it is ok to just let users to take care of the shapes, just as PyTorch. The dynamic graph by...

Agree. I think one important thing is figuring out what functionality a Module should have (for example, load(), save(), fit(), predict() ...), and how these could be implemented under the...