[RLlib] Examples folder restructuring (Model examples; final part). (#8278)

- This PR completes any previously missing PyTorch Model counterparts to TFModels in examples/models.
- It also makes sure, all example scripts in the rllib/examples folder are tested for both frameworks and learn the given task (this is often currently not checked) using a --as-test flag in connection with a --stop-reward.

rllib/examples/custom_env.py

@@ -7,20 +7,32 @@ This example shows:
 
 You can visualize experiment results in ~/ray_results using TensorBoard.
 """
-
+import argparse
 import numpy as np
 import gym
-from ray.rllib.models import ModelCatalog
-from ray.rllib.models.tf.tf_modelv2 import TFModelV2
-from ray.rllib.models.tf.fcnet_v2 import FullyConnectedNetwork
 from gym.spaces import Discrete, Box
 
 import ray
 from ray import tune
-from ray.rllib.utils import try_import_tf
 from ray.tune import grid_search
+from ray.rllib.models import ModelCatalog
+from ray.rllib.models.tf.tf_modelv2 import TFModelV2
+from ray.rllib.models.tf.fcnet_v2 import FullyConnectedNetwork
+from ray.rllib.models.torch.torch_modelv2 import TorchModelV2
+from ray.rllib.models.torch.fcnet import FullyConnectedNetwork as TorchFC
+from ray.rllib.utils.framework import try_import_tf, try_import_torch
+from ray.rllib.utils.test_utils import check_learning_achieved
 
 tf = try_import_tf()
+torch, nn = try_import_torch()
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--run", type=str, default="PPO")
+parser.add_argument("--torch", action="store_true")
+parser.add_argument("--as-test", action="store_true")
+parser.add_argument("--stop-iters", type=int, default=50)
+parser.add_argument("--stop-timesteps", type=int, default=100000)
+parser.add_argument("--stop-reward", type=float, default=0.1)
 
 
 class SimpleCorridor(gym.Env):
@@ -46,11 +58,11 @@ class SimpleCorridor(gym.Env):
         elif action == 1:
             self.cur_pos += 1
         done = self.cur_pos >= self.end_pos
-        return [self.cur_pos], 1 if done else 0, done, {}
+        return [self.cur_pos], 1.0 if done else -0.1, done, {}
 
 
 class CustomModel(TFModelV2):
-    """Example of a custom model that just delegates to a fc-net."""
+    """Example of a keras custom model that just delegates to an fc-net."""
 
     def __init__(self, obs_space, action_space, num_outputs, model_config,
                  name):
@@ -67,26 +79,58 @@ class CustomModel(TFModelV2):
         return self.model.value_function()
 
 
+class TorchCustomModel(TorchModelV2, nn.Module):
+    """Example of a PyTorch custom model that just delegates to a fc-net."""
+
+    def __init__(self, obs_space, action_space, num_outputs, model_config,
+                 name):
+        TorchModelV2.__init__(self, obs_space, action_space, num_outputs,
+                              model_config, name)
+        nn.Module.__init__(self)
+
+        self.torch_sub_model = TorchFC(obs_space, action_space, num_outputs,
+                                       model_config, name)
+
+    def forward(self, input_dict, state, seq_lens):
+        input_dict["obs"] = input_dict["obs"].float()
+        fc_out, _ = self.torch_sub_model(input_dict, state, seq_lens)
+        return fc_out, []
+
+    def value_function(self):
+        return torch.reshape(self.torch_sub_model.value_function(), [-1])
+
+
 if __name__ == "__main__":
+    args = parser.parse_args()
+    ray.init()
+
     # Can also register the env creator function explicitly with:
     # register_env("corridor", lambda config: SimpleCorridor(config))
-    ray.init()
-    ModelCatalog.register_custom_model("my_model", CustomModel)
-    tune.run(
-        "PPO",
-        stop={
-            "timesteps_total": 10000,
+    ModelCatalog.register_custom_model(
+        "my_model", TorchCustomModel if args.torch else CustomModel)
+
+    config = {
+        "env": SimpleCorridor,  # or "corridor" if registered above
+        "env_config": {
+            "corridor_length": 5,
         },
-        config={
-            "env": SimpleCorridor,  # or "corridor" if registered above
-            "model": {
-                "custom_model": "my_model",
-            },
-            "vf_share_layers": True,
-            "lr": grid_search([1e-2, 1e-4, 1e-6]),  # try different lrs
-            "num_workers": 1,  # parallelism
-            "env_config": {
-                "corridor_length": 5,
-            },
+        "model": {
+            "custom_model": "my_model",
         },
-    )
+        "vf_share_layers": True,
+        "lr": grid_search([1e-2, 1e-4, 1e-6]),  # try different lrs
+        "num_workers": 1,  # parallelism
+        "use_pytorch": args.torch
+    }
+
+    stop = {
+        "training_iteration": args.stop_iters,
+        "timesteps_total": args.stop_timesteps,
+        "episode_reward_mean": args.stop_reward,
+    }
+
+    results = tune.run(args.run, config=config, stop=stop)
+
+    if args.as_test:
+        check_learning_achieved(results, args.stop_reward)
+    ray.shutdown()