Hello Maintainer,
I've noticed a duplicate agent_grad_norm in the maddpg_learner.py file that appears to be redundant and should be removed. The duplication can be found in the following lines:
|
if t_env - self.log_stats_t >= self.args.learner_log_interval: |
|
self.logger.log_stat("critic_loss", loss.item(), t_env) |
|
self.logger.log_stat("critic_grad_norm", critic_grad_norm.item(), t_env) |
|
self.logger.log_stat("agent_grad_norm", agent_grad_norm.item(), t_env) |
|
mask_elems = mask.sum().item() |
|
self.logger.log_stat("td_error_abs", masked_td_error.abs().sum().item() / mask_elems, t_env) |
|
self.logger.log_stat("q_taken_mean", (q_taken).sum().item() / mask_elems, t_env) |
|
self.logger.log_stat("target_mean", targets.sum().item() / mask_elems, t_env) |
|
self.logger.log_stat("pg_loss", pg_loss.item(), t_env) |
|
self.logger.log_stat("agent_grad_norm", agent_grad_norm, t_env) |
|
self.log_stats_t = t_env |
Additionally, I saw that there has been a pull request submitted which addresses this issue: #60
Best regards,
Maonan Wang
Hello Maintainer,
I've noticed a duplicate
agent_grad_normin themaddpg_learner.pyfile that appears to be redundant and should be removed. The duplication can be found in the following lines:epymarl/src/learners/maddpg_learner.py
Lines 136 to 146 in 4b743a8
Additionally, I saw that there has been a pull request submitted which addresses this issue: #60
Best regards,
Maonan Wang