learning_by_grad_by_grad_repro

learning_by_grad_by_grad_repro copied to clipboard

Hi... Nice work. I was trying to run the code. While running the part where all optimizers are considered, I got the error "Optimizers got an empty parameter list. This...

BasitAlawode

Hi, I tried applying the MNIST code to FashionMNIST however the code does not seem to train the LSMT optimizer properly. To change to the FashionMNIST dataset, all I changed...

thomasarmstrong98

Thank you for the great effort! 1. Is the reason of not using nn.Parameters but using Variables can be explained by [this post](https://discuss.pytorch.org/t/nn-parameter-doesnt-retain-grad-fn/29214)? 2. I think relying on Variables is...

chenwydj

Code only supports toy examples and can be cleaned up by implementing zero_grad

1

Currently, we recreate the optimizee each time, instead of doing this, we can simply use the `zero_grad()` code to reset the gradients in the optimizee.

alexvicegrab

offset variable has not been updated

1

hi, I found your notebook really of great value to me. But after reading and running your code, I got several questions afflicting me for a while. 1. In `do_fit`...

itongworld