leezu commented on a change in pull request #16900: introduce gradient update
handler to the base estimator
URL: https://github.com/apache/incubator-mxnet/pull/16900#discussion_r351075164
##########
File path: python/mxnet/gluon/contrib/estimator/event_handler.py
##########
@@ -706,3 +714,29 @@ def train_end(self, estimator, *args, **kwargs):
estimator.logger.info('[Epoch %d] EarlyStoppingHanlder: '
'early stopping due to %s not improving',
self.stopped_epoch, self.monitor.get()[0])
+
+class GradientUpdateHandler(BatchEnd):
+ """Gradient Update Handler that apply gradients on network weights
+
+ :py:class:`GradientUpdateHandler` takes the priority level. It updates
weight parameters
+ at the end of each batch
+
+ Parameters
+ ----------
+ priority : scalar, default -np.Inf
Review comment:
Should a integer be used instead of `-np.Inf`? Maybe there are some
use-cases where a handler should be run prior to the gradient handler?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to 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
