Morvan
Morvan
https://github.com/MorvanZhou/PyTorch-Tutorial/blob/master/tutorial-contents/406_GAN.py#L52-L53 ``` opt_D = torch.optim.Adam(D.parameters(), lr=LR_D) opt_G = torch.optim.Adam(G.parameters(), lr=LR_G) ``` This will update G and D separately.
Is the bug related to pytorch's version? The pytorch '0.4.1' runs fine.
Get the same issue on MacOS 12.3.1
I have updated the code for tf 1.8.0. It works fine when I test it with 8 workers.
This is from a3c without rnn. The a3c with rnn may have some issues. 
The `lnet.load_state_dict()` function shows as below: ```python def load_state_dict(self, state_dict): # deepcopy, to be consistent with module API state_dict = deepcopy(state_dict) # Validate the state_dict groups = self.param_groups saved_groups =...
> It will be very kind of you to explain if there might be conficts between workers without locking the shared model. A lock could be applied in this case,...
Can you fix the random seed and test the result again? Fix random seed by set: ``` tf.set_random_seed(1) np.random.seed(1) ```
I did one experiment about "Pendulum", it works good. But when I increase the action dimension, like 'BipedalWalker-v2', it's action variance jumps too high and not converges. I till cannot...
@naiba 哈哈，都是缘分