eric-haibin-lin opened a new issue #13752: Adam, AdaMax and FTML cannot be used with Trainer(update_on_kv=False) URL: https://github.com/apache/incubator-mxnet/issues/13752 These optimizers adjust the learning rate by the number of optimization steps, which is problematic if `update_on_kv=False` is set and multiple GPUs share the same optimizer object (leading to incorrect count of optimization steps). For example: https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/optimizer/optimizer.py#L1093 cc @szha @sandeep-krishnamurthy
---------------------------------------------------------------- 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
