Skip to content

Minor changes for torch > 1.4 compatibility #13

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

Open
dssrgu wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Minor changes for torch > 1.4 compatibility #13

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
14 changes: 7 additions & 7 deletions d4rl/rlkit/torch/sac/cql.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,9 +142,9 @@ def _get_policy_actions(self, obs, num_actions, network=None):
obs_temp, reparameterize=True, return_log_prob=True,
)
if not self.discrete:
return new_obs_actions, new_obs_log_pi.view(obs.shape[0], num_actions, 1)
return new_obs_actions.detach(), new_obs_log_pi.view(obs.shape[0], num_actions, 1).detach()
else:
return new_obs_actions
return new_obs_actions.detach()

def train_from_torch(self, batch):
self._current_epoch += 1
Expand Down Expand Up @@ -283,6 +283,11 @@ def train_from_torch(self, batch):
"""
Update networks
"""
self._num_policy_update_steps += 1
self.policy_optimizer.zero_grad()
policy_loss.backward(retain_graph=False)
self.policy_optimizer.step()

# Update the Q-functions iff
self._num_q_update_steps += 1
self.qf1_optimizer.zero_grad()
Expand All @@ -294,11 +299,6 @@ def train_from_torch(self, batch):
qf2_loss.backward(retain_graph=True)
self.qf2_optimizer.step()

self._num_policy_update_steps += 1
self.policy_optimizer.zero_grad()
policy_loss.backward(retain_graph=False)
self.policy_optimizer.step()

"""
Soft Updates
"""
Expand Down