Add training on custom openpi datasets

Cleanup instructions

clean up doc

pass linter

updates

Add test

scripts/serve_policy.py

@@ -16,6 +16,7 @@ from openpi.policies import libero_policy
 from openpi.policies import policy as _policy
 from openpi.policies import policy_config as _policy_config
 from openpi.serving import websocket_policy_server
+from openpi.shared import delta_actions
 from openpi.training import config as _config
 
 
@@ -146,7 +147,7 @@ def create_default_policy(
     logging.info("Creating policy...")
     match env:
         case EnvMode.ALOHA:
-            delta_action_mask = _policy_config.make_bool_mask(6, -1, 6, -1)
+            delta_action_mask = delta_actions.make_bool_mask(6, -1, 6, -1)
             config = make_policy_config(
                 input_layers=[
                     aloha_policy.ActInputsRepack(),