Skip to content

Fix sync_target #28

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

Merged
sdpkjc merged 2 commits into from
Dec 2, 2022
Merged

Fix sync_target #28

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ed73c27
fix sync_target
sdpkjc Dec 2, 2022
bac80b4
Merge branch 'main' into fix/sync_target
sdpkjc Dec 2, 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
5 changes: 3 additions & 2 deletions abcdrl/ddpg.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -227,7 +227,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
)
act_ts += noise
act_np = act_ts.cpu().numpy().clip(self.kwargs["act_space"].low, self.kwargs["act_space"].high)

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["policy_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -242,8 +245,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts, self.sample_step % self.kwargs["policy_frequency"] == 0)
if self.sample_step % self.kwargs["policy_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
return log_data

Expand Down
5 changes: 3 additions & 2 deletions abcdrl/ddqn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,7 +172,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
with torch.no_grad():
_, act_ts = self.alg.predict(obs_ts).max(dim=1)
act_np = act_ts.cpu().numpy()

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -187,8 +190,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts)
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
log_data["epsilon"] = self._get_epsilon()
return log_data
Expand Down
5 changes: 3 additions & 2 deletions abcdrl/dqn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -170,7 +170,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
with torch.no_grad():
_, act_ts = self.alg.predict(obs_ts).max(dim=1)
act_np = act_ts.cpu().numpy()

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -185,8 +188,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts)
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
log_data["epsilon"] = self._get_epsilon()
return log_data
Expand Down
5 changes: 3 additions & 2 deletions abcdrl/pdqn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -281,7 +281,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
with torch.no_grad():
_, act_ts = self.alg.predict(obs_ts).max(dim=1)
act_np = act_ts.cpu().numpy()

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: PrioritizedReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -296,8 +299,6 @@ def learn(self, data: PrioritizedReplayBuffer.Samples[np.ndarray]) -> dict[str,
)

log_data = self.alg.learn(data_ts)
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
log_data["epsilon"] = self._get_epsilon()
return log_data
Expand Down
5 changes: 3 additions & 2 deletions abcdrl/sac.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -274,7 +274,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
with torch.no_grad():
act_ts = self.alg.predict(obs_ts)
act_np = act_ts.cpu().numpy()

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -289,8 +292,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts, self.sample_step % self.kwargs["policy_frequency"] == 0)
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
return log_data

Expand Down
5 changes: 3 additions & 2 deletions abcdrl/td3.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -253,7 +253,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
)
act_ts += noise
act_np = act_ts.cpu().numpy().clip(self.kwargs["act_space"].low, self.kwargs["act_space"].high)

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["policy_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -268,8 +271,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts, self.sample_step % self.kwargs["policy_frequency"] == 0)
if self.sample_step % self.kwargs["policy_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
return log_data

Expand Down
5 changes: 3 additions & 2 deletions abcdrl_copy_from/dqn_all_wrappers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,7 +171,10 @@ def sample(self, obs: np.ndarray) -> np.ndarray:
with torch.no_grad():
_, act_ts = self.alg.predict(obs_ts).max(dim=1)
act_np = act_ts.cpu().numpy()

self.sample_step += self.kwargs["num_envs"]
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
return act_np

def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
Expand All @@ -186,8 +189,6 @@ def learn(self, data: ReplayBuffer.Samples[np.ndarray]) -> dict[str, Any]:
)

log_data = self.alg.learn(data_ts)
if self.sample_step % self.kwargs["target_network_frequency"] == 0:
self.alg.sync_target()
self.learn_step += 1
log_data["epsilon"] = self._get_epsilon()
return log_data
Expand Down