wkcn opened a new issue #9307: How to clear the optimizer states (reset a 
optimizer)
URL: https://github.com/apache/incubator-mxnet/issues/9307
 
 
   Hi, there.
   
   I wrote an `eval_end_callback` function, and there is a condition that 
changing the learning rate of the optimizer when a metric decreases.
   
   In mxnet.optimizer.SGD, there are the optimizer states.
   
   ```python 
   rescaled_grad = lr * rescale_grad * clip(grad, clip_gradient) + wd * weight
   state = momentum * state + rescaled_grad
   weight = weight - state
   ```
   
   I want to set the state value zero as the same as the initialization of the 
optimizer.
   
   How to clear the optimizer states?
   
   Thank you!

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to