Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix action scaling bug in sac #591

Merged
merged 3 commits into from
Apr 11, 2022
Merged

Fix action scaling bug in sac #591

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3af843a
Fix action scaling bug in sac
ChenDRAG Apr 11, 2022
bfaf9f0
fix
ChenDRAG Apr 11, 2022
aee24fb
fix ci
Trinkle23897 Apr 11, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 3 additions & 9 deletions tianshou/policy/modelfree/sac.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
import torch
from torch.distributions import Independent, Normal

from tianshou.data import Batch, ReplayBuffer, to_torch_as
from tianshou.data import Batch, ReplayBuffer
from tianshou.exploration import BaseNoise
from tianshou.policy import DDPGPolicy

Expand Down Expand Up @@ -121,15 +121,9 @@ def forward( # type: ignore
# apply correction for Tanh squashing when computing logprob from Gaussian
# You can check out the original SAC paper (arXiv 1801.01290): Eq 21.
# in appendix C to get some understanding of this equation.
if self.action_scaling and self.action_space is not None:
low, high = self.action_space.low, self.action_space.high # type: ignore
action_scale = to_torch_as((high - low) / 2.0, act)
else:
action_scale = 1.0 # type: ignore
squashed_action = torch.tanh(act)
log_prob = log_prob - torch.log(
action_scale * (1 - squashed_action.pow(2)) + self.__eps
).sum(-1, keepdim=True)
log_prob = log_prob - torch.log((1 - squashed_action.pow(2)) +
self.__eps).sum(-1, keepdim=True)
return Batch(
logits=logits,
act=squashed_action,
Expand Down