Le Zhang

in lines 240-241 of `core / reanalyze_worker.py`, try changing them to ``` trained_steps = ray.get(self.storage.get_counter.remote()) target_weights = None ``` and changing lines 252-253 to ``` if new_model_index > self.last_model_index: self.last_model_index...

btw, in train/mean_score of your posted plot, 100K in x-axis is not for Atari 100K, but for Atrai 10M (i.e., 10M interactions with the env)? is the understanding above right?

did not try the exp in the large scale as you discussed. but the change on codes relevant to `target_weights` makes the `train.sh` be runnable. and decreasing the `gpu_actor` really...

no. it is just because this would save RAM memory, so `train.sh` would run without breaking until the end.

+1 error also for SpaceInvaders + SAC_Discrete. So this now makes things harder to be verified for the results in the SAC_Discrete arXiv preprint paper. (Paper: [Soft Actor-Critic for Discrete...

> Thanks for your interest. I reviewed this part of the code carefully, it is about parent node update. These codes are used to implement the functionality where all the...

+1 Would appreciate if the code for computing DCS (esp. how the biophysical and statistical properties are implemented in the form of code when computing DCS) is given (rather than...