Adding test configurations

experiments/tests/MARWIL.yaml

+flatland-random-sparse-small-tree-marwil-fc-ppo:
+    run: MARWIL
+    env: flatland_random_sparse_small
+    stop:
+        timesteps_total: 100000  # 1e7
+    checkpoint_freq: 10
+    checkpoint_at_end: True
+    keep_checkpoints_num: 5
+    checkpoint_score_attr: episode_reward_mean
+    config:
+        beta: 0
+        # grid_search: [0, 1]   # compare IL (beta=0) vs MARWIL
+        input: /tmp/flatland-out
+        # effective batch_size: train_batch_size * num_agents_in_each_environment [5, 10]
+        # see https://github.com/ray-project/ray/issues/4628
+        train_batch_size: 1000  # 5000
+        rollout_fragment_length: 50  # 100
+        num_workers: 2
+        num_envs_per_worker: 1
+        batch_mode: truncate_episodes
+        observation_filter: NoFilter
+        num_gpus: 0
+
+        env_config:
+            min_seed: 1002
+            max_seed: 213783
+            min_test_seed: 0
+            max_test_seed: 100
+            # After how many episodes the level should be regenerated:
+            reset_env_freq: 1
+            observation: tree
+            observation_config:
+                max_depth: 2
+                shortest_path_max_depth: 30
+            regenerate_rail_on_reset: True
+            regenerate_schedule_on_reset: True
+            render: False
+
+        model:
+            fcnet_activation: relu
+            fcnet_hiddens: [256, 256]
+            vf_share_layers: True  # False
+
+        # evaluation_num_workers: 1
+        # # Enable evaluation, once per training iteration.
+        # evaluation_interval: 25
+        # # Run 1 episode each time evaluation runs.
+        # evaluation_num_episodes: 1
+        # # Override the env config for evaluation.
+        # evaluation_config:
+        #     explore: False
+        #     env_config:
+        #         min_seed: 1002
+        #         max_seed: 213783
+        #         min_test_seed: 0
+        #         max_test_seed: 100
+        #         # After how many episodes the level should be regenerated:
+        #         reset_env_freq: 1
+        #         observation: tree
+        #         observation_config:
+        #             max_depth: 2
+        #             shortest_path_max_depth: 30
+        #         regenerate_rail_on_reset: True
+        #         regenerate_schedule_on_reset: True
+        #         render: False

experiments/tests/conflict_obs_apex.yaml

+flatland-random-sparse-small-tree-fc-ppo:
+    run: APEX
+    env: flatland_sparse
+    stop:
+        timesteps_total: 50000
+    checkpoint_freq: 10
+    checkpoint_at_end: True
+    keep_checkpoints_num: 5
+    checkpoint_score_attr: episode_reward_mean
+    config:
+        num_workers: 2
+        num_envs_per_worker: 5
+        num_gpus: 0
+
+        env_config:
+            observation: localConflict
+            observation_config:
+                max_depth: 2
+                shortest_path_max_depth: 30
+                n_local: 5
+
+            generator: sparse_rail_generator
+            generator_config: small_v0
+
+            wandb:
+                project: flatland
+                entity: masterscrat
+                tags: ["small_v0", "local_conflict_obs", "apex"] # TODO should be set programmatically
+
+        model:
+            fcnet_activation: relu
+            fcnet_hiddens: [256, 256]
+            vf_share_layers: True  # False

experiments/tests/global_obs_ppo.yaml

+flatland-sparse-global-conv-ppo:
+    run: PPO
+    env: flatland_sparse
+    stop:
+        timesteps_total: 500  # 1e7
+    checkpoint_freq: 10
+    checkpoint_at_end: True
+    keep_checkpoints_num: 5
+    checkpoint_score_attr: episode_reward_mean
+    config:
+        clip_rewards: True
+        clip_param: 0.1
+        vf_clip_param: 500.0
+        entropy_coeff: 0.01
+        # effective batch_size: train_batch_size * num_agents_in_each_environment [5, 10]
+        # see https://github.com/ray-project/ray/issues/4628
+        train_batch_size: 1000  # 5000
+        rollout_fragment_length: 50  # 100
+        sgd_minibatch_size: 100  # 500
+        num_sgd_iter: 10
+        num_workers: 2
+        num_envs_per_worker: 2
+        batch_mode: truncate_episodes
+        observation_filter: NoFilter
+        vf_share_layers: True
+        vf_loss_coeff: 0.5
+        num_gpus: 0
+
+        env_config:
+            observation: global
+            observation_config:
+                max_width: 32
+                max_height: 32
+
+            generator: sparse_rail_generator
+            generator_config: small_v0
+
+            wandb:
+                project: flatland
+                entity: masterscrat
+                tags: ["small_v0", "global_obs"] # TODO should be set programmatically
+
+        model:
+            custom_model: global_obs_model
+            custom_options:
+                architecture: impala
+                architecture_options:
+                    residual_layers: [[16, 2], [32, 4]]

test.sh

+echo "===================="
+echo "MARWIL"
+echo "===================="
+time python ./train.py -f experiments/tests/MARWIL.yaml
+
+echo "===================="
+echo "GLOBAL OBS"
+echo "===================="
+time python ./train.py -f experiments/tests/global_obs.yaml
\ No newline at end of file