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

Sim Library Improvements #116

Merged
merged 6 commits into from
Nov 21, 2024
+39 −13
Merged

Sim Library Improvements #116

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
cae7a62
add cmd arg for h5 output path with sensible default
alik-git Nov 20, 2024
833f03a
save current actions to h5
alik-git Nov 20, 2024
6a65d01
add h5out as arg to produce_sim_data as well
alik-git Nov 20, 2024
df09d0d
compute actions before saving them
alik-git Nov 20, 2024
102a770
typo fix: args.log_h5 -> log_h5
alik-git Nov 20, 2024
0eb70d3
fix: rename actions to prev_actions, and set prev_action to curr_acti…
alik-git Nov 21, 2024
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
6 changes: 4 additions & 2 deletions sim/h5_logger.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@ def _create_h5_file(self):
h5_file = h5py.File(h5_file_path, "w")

# Create datasets for logging actions and observations
dset_actions = h5_file.create_dataset("prev_actions", (self.max_timesteps, self.num_actions), dtype=np.float32)
dset_prev_actions = h5_file.create_dataset("prev_actions", (self.max_timesteps, self.num_actions), dtype=np.float32)
dset_2D_command = h5_file.create_dataset("observations/2D_command", (self.max_timesteps, 2), dtype=np.float32)
dset_3D_command = h5_file.create_dataset("observations/3D_command", (self.max_timesteps, 3), dtype=np.float32)
dset_q = h5_file.create_dataset("observations/q", (self.max_timesteps, self.num_actions), dtype=np.float32)
Expand All @@ -42,10 +42,12 @@ def _create_h5_file(self):
dset_euler = h5_file.create_dataset("observations/euler", (self.max_timesteps, 3), dtype=np.float32)
dset_t = h5_file.create_dataset("observations/t", (self.max_timesteps, 1), dtype=np.float32)
dset_buffer = h5_file.create_dataset("observations/buffer", (self.max_timesteps, self.num_observations), dtype=np.float32)
dset_curr_actions = h5_file.create_dataset("curr_actions", (self.max_timesteps, self.num_actions), dtype=np.float32)

# Map datasets for easy access
h5_dict = {
"prev_actions": dset_actions,
"prev_actions": dset_prev_actions,
"curr_actions": dset_curr_actions,
"2D_command": dset_2D_command,
"3D_command": dset_3D_command,
"joint_pos": dset_q,
Expand Down
1 change: 1 addition & 0 deletions sim/sim2sim.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -252,6 +252,7 @@ def run_mujoco(
"joint_pos": cur_pos_obs.astype(np.float32),
"joint_vel": cur_vel_obs.astype(np.float32),
"prev_actions": actions.astype(np.float32),
"curr_actions": target_q.astype(np.float32),
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

current actions are actions listed below

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

addressed in df09d0d

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is wrong, you are now saving latest actions and positions (target_q which are scaled actions).

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

okay should be fixed in 0eb70d3

i renamed actions to prev_actions since that is what it is being used as, and use curr_actions for the output of the model. at the end of the inner loop iteration, I set prev_actions = curr_actions.

"ang_vel": omega.astype(np.float32),
"euler_rotation": eu_ang.astype(np.float32),
"buffer": hist_obs.astype(np.float32)
Expand Down
Loading