WIP

lerobot/configs/policy/tdmpc.yaml

@@ -1,6 +1,7 @@
 # @package _global_
 
 seed: 1
+dataset_repo_id: lerobot/xarm_lift_medium_replay
 
 training:
   offline_steps: 25000

tests/scripts/save_policy_to_safetensor.py

@@ -82,9 +82,9 @@ def save_policy_to_safetensors(output_dir, env_name, policy_name):
 
 if __name__ == "__main__":
     env_policies = [
-        # ("xarm", "tdmpc"),
-        ("pusht", "diffusion"),
-        ("aloha", "act"),
+        ("xarm", "tdmpc"),
+        # ("pusht", "diffusion"),
+        # ("aloha", "act"),
     ]
     for env, policy in env_policies:
         save_policy_to_safetensors("tests/data/save_policy_to_safetensors", env, policy)

tests/test_policies.py

@@ -239,7 +239,7 @@ def test_normalize(insert_temporal_dim):
 @pytest.mark.parametrize(
     "env_name,policy_name",
     [
-        # ("xarm", "tdmpc"),
+        ("xarm", "tdmpc"),
         ("pusht", "diffusion"),
         ("aloha", "act"),
     ],