support zero-dim tensor for paddle future version

PaddlePaddle / PARL

A high-performance distributed training framework for Reinforcement Learning

https://parl.readthedocs.io/

Apache License 2.0

3.22k stars 816 forks source link

support zero-dim tensor for paddle future version #980

Closed DimariaW closed 1 year ago

DimariaW commented 1 year ago

change 'action.argmax().numpy()[0]' to 'int(action.argmax())' change 'loss.numpy()[0]' to 'float(loss)' test changes in carpole and work

CLAassistant commented 1 year ago

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

dimariaw seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}