refactor vis_encoder_type and add to doc

refactor vis_encoder_type and add to doc

docs/Training-PPO.md

 
 Typical Range: `32` - `512`
 
+### (Optional) Visual Encoder Type
+
+`vis_encode_type` corresponds to the encoder type for encoding visual observations.
+Valid options include:
+* `simple` (default): a simple encoder which consists of two convolutional layers
+* `nature_cnn`: CNN implementation proposed by Mnih et al.(https://www.nature.com/articles/nature14236), 
+consisting of three convolutional layers
+* `resnet`: IMPALA Resnet implementation (https://arxiv.org/abs/1802.01561),
+consisting of three stacked layers, each with two risidual blocks, making a 
+much larger network than the other two.
+
+Options: `simple`, `nature_cnn`, `resnet`
+
 ## (Optional) Recurrent Neural Network Hyperparameters
 
 The below hyperparameters are only used when `use_recurrent` is set to true.

ml-agents/mlagents/trainers/models.py

 
 
 class EncoderType(Enum):
-    RESNET = "resnet"
+    SIMPLE = "simple"
-    DEFAUL = "default"
+    RESNET = "resnet"
 
 
 class LearningModel(object):
         :param reuse: Whether to re-use the weights within the same scope.
         :return: List of hidden layer tensors.
         """
-        print("creating nature cnn")
         with tf.variable_scope(scope):
             conv1 = tf.layers.conv2d(
                 image_input,
         :param reuse: Whether to re-use the weights within the same scope.
         :return: List of hidden layer tensors.
         """
-        print("creating resnet")
         n_channels = [16, 32, 32]  # channel for each stack
         n_blocks = 2  # number of residual blocks
         with tf.variable_scope(scope):
         )
 
     def create_observation_streams(
-        self, num_streams, h_size, num_layers, vis_encode_type="default"
-    ):
+        self,
+        num_streams: int,
+        h_size: int,
+        num_layers: int,
+        vis_encode_type: EncoderType = EncoderType.SIMPLE,
+    ) -> tf.Tensor:
         """
         Creates encoding stream for observations.
         :param num_streams: Number of streams to create.
             visual_encoders = []
             hidden_state, hidden_visual = None, None
             if self.vis_obs_size > 0:
-                vis_encode_type = EncoderType(vis_encode_type)
                 if vis_encode_type == EncoderType.RESNET:
                     for j in range(brain.number_visual_observations):
                         encoded_visual = self.create_resnet_visual_observation_encoder(
             self.value_heads[name] = value
         self.value = tf.reduce_mean(list(self.value_heads.values()), 0)
 
-    def create_cc_actor_critic(self, h_size, num_layers, vis_encode_type):
+    def create_cc_actor_critic(
+        self, h_size: int, num_layers: int, vis_encode_type: EncoderType
+    ) -> None:
         """
         Creates Continuous control actor-critic model.
         :param h_size: Size of hidden linear layers.
             (tf.identity(self.all_old_log_probs)), axis=1, keepdims=True
         )
 
-    def create_dc_actor_critic(self, h_size, num_layers, vis_encode_type):
+    def create_dc_actor_critic(
+        self, h_size: int, num_layers: int, vis_encode_type: EncoderType
+    ) -> None:
         """
         Creates Discrete control actor-critic model.
         :param h_size: Size of hidden linear layers.

ml-agents/mlagents/trainers/ppo/models.py

 import numpy as np
 
 import tensorflow as tf
-from mlagents.trainers.models import LearningModel
+from mlagents.trainers.models import LearningModel, EncoderType
 
 logger = logging.getLogger("mlagents.trainers")
 
         m_size=None,
         seed=0,
         stream_names=None,
-        vis_encode_type="default",
+        vis_encode_type=EncoderType.SIMPLE,
     ):
         """
         Takes a Unity environment and model-specific hyper-parameters and returns the

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

 
 from mlagents.envs.timers import timed
 from mlagents.trainers import BrainInfo, ActionInfo
+from mlagents.trainers.models import EncoderType
 from mlagents.trainers.ppo.models import PPOModel
 from mlagents.trainers.tf_policy import TFPolicy
 from mlagents.trainers.components.reward_signals.reward_signal_factory import (
                 m_size=self.m_size,
                 seed=seed,
                 stream_names=list(reward_signal_configs.keys()),
-                vis_encode_type=trainer_params["vis_encode_type"],
+                vis_encode_type=EncoderType(
+                    trainer_params.get("vis_encode_type", "simple")
+                ),
             )
             self.model.create_ppo_optimizer()
 

ml-agents/mlagents/trainers/ppo/trainer.py

             "memory_size",
             "model_path",
             "reward_signals",
-            "vis_encode_type",
         ]
         self.check_param_keys()
 

ml-agents/mlagents/trainers/tests/test_bcmodule.py

         sequence_length: 64
         summary_freq: 1000
         use_recurrent: false
-        vis_encode_type: default
         memory_size: 8
         pretraining:
           demo_path: ./demos/ExpertPyramid.demo

ml-agents/mlagents/trainers/tests/test_environments/test_simple.py

                 extrinsic:
                     strength: 1.0
                     gamma: 0.99
-            vis_encode_type: default
     """
     # Create controller and begin training.
     with tempfile.TemporaryDirectory() as dir:

ml-agents/mlagents/trainers/tests/test_ppo.py

         sequence_length: 64
         summary_freq: 1000
         use_recurrent: false
-        vis_encode_type: default
         memory_size: 8
         curiosity_strength: 0.0
         curiosity_enc_size: 1
         "sequence_length": 64,
         "summary_freq": 3000,
         "use_recurrent": False,
-        "vis_encode_type": "default",
         "use_curiosity": False,
         "curiosity_strength": 0.01,
         "curiosity_enc_size": 128,

ml-agents/mlagents/trainers/tests/test_reward_signals.py

         sequence_length: 64
         summary_freq: 1000
         use_recurrent: false
-        vis_encode_type: default
         memory_size: 8
         curiosity_strength: 0.0
         curiosity_enc_size: 1