1 layer everything

config/sac_transfer/3DBallCloud.yaml

       save_replay_buffer: false
       init_entcoef: 0.5
       reward_signal_steps_per_update: 10.0
-      encoder_layers: 2
+      encoder_layers: 1
-      forward_layers: 0
+      forward_layers: 1
       value_layers: 1
       feature_size: 16
       # separate_value_net: true

config/sac_transfer/3DBallHardCloud.yaml

       save_replay_buffer: false
       init_entcoef: 1.0
       reward_signal_steps_per_update: 10.0
-      encoder_layers: 2
+      encoder_layers: 1
-      forward_layers: 0
+      forward_layers: 1
       value_layers: 1
       feature_size: 16
       # separate_value_net: true

config/sac_transfer/3DBallHardTransferCloud.yaml

       save_replay_buffer: false
       init_entcoef: 1.0
       reward_signal_steps_per_update: 10.0
-      encoder_layers: 2
+      encoder_layers: 1
-      forward_layers: 0
+      forward_layers: 1
       value_layers: 1
       feature_size: 16
       # separate_value_net: true
       use_transfer: true
       load_model: true
       train_model: false
-      transfer_path: "results/sac-ball-f16-p1f0/3DBall"
+      transfer_path: "results/sac-ball-f16-e1p1f1/3DBall"
     network_settings:
       normalize: true
       hidden_units: 128

ml-agents/mlagents/trainers/policy/transfer_policy.py

             #     reuse_encoder,
             # )
 
-            # self.action_encoder = self._create_action_encoder(
-            #     self.current_action,
-            #     self.h_size,
-            #     self.action_feature_size,
-            #     action_layers
-            # )
+            self.action_encoder = self._create_action_encoder(
+                self.current_action,
+                self.h_size,
+                self.action_feature_size,
+                action_layers
+            )
 
             if self.inverse_model:
                 with tf.variable_scope("inverse"):
 
                 self.predict, self.predict_distribution = self.create_forward_model(
                     self.encoder,
-                    self.current_action,
+                    self.action_encoder,
                     forward_layers,
                     var_predict=var_predict,
                     separate_train=separate_model_train
                     self.targ_encoder,
-                    self.current_action,
+                    self.action_encoder,
                     forward_layers,
                     var_predict=var_predict,
                     reuse=True,
             if predict_return:
                 with tf.variable_scope("reward"):
                     self.create_reward_model(
-                        self.encoder, self.current_action, forward_layers, separate_train=separate_model_train
+                        self.encoder, self.action_encoder, forward_layers, separate_train=separate_model_train
                     )
 
             if self.use_bisim:
         num_layers: int,
         reuse: bool=False
     ) -> tf.Tensor:
+
+        if num_layers < 0:
+            return self.current_action
         
         hidden_stream = ModelUtils.create_vector_observation_encoder(
             action, 
             encoding_checkpoint = os.path.join(self.model_path, f"encoding.ckpt")
             encoding_saver.save(self.sess, encoding_checkpoint)
 
-            # action_vars = tf.get_collection(
-            #     tf.GraphKeys.TRAINABLE_VARIABLES, "action_enc"
-            # )
-            # action_saver = tf.train.Saver(action_vars)
-            # action_checkpoint = os.path.join(self.model_path, f"action_enc.ckpt")
-            # action_saver.save(self.sess, action_checkpoint)
+            action_vars = tf.get_collection(
+                tf.GraphKeys.TRAINABLE_VARIABLES, "action_enc"
+            )
+            if len(action_vars) > 0:
+                action_saver = tf.train.Saver(action_vars)
+                action_checkpoint = os.path.join(self.model_path, f"action_enc.ckpt")
+                action_saver.save(self.sess, action_checkpoint)
 
             latent_vars = tf.get_collection(
                 tf.GraphKeys.TRAINABLE_VARIABLES, "encoding/latent"

ml-agents/mlagents/trainers/settings.py

 
     # Network
     encoder_layers: int = 1
-    action_layers: int = 1
+    action_layers: int = -1
     policy_layers: int = 1
     value_layers: int = 1
     forward_layers: int = 1