Added medium stochastic env and experiment

envs/flatland/generator_configs/medium_stoch_v0.yaml

+# taken from Flatland examples/flatland_2_0_example.py
+width: 100
+height: 100
+number_of_agents: 100
+max_num_cities: 30
+grid_mode: False
+max_rails_between_cities: 2
+max_rails_in_city: 8
+malfunction_rate: 30
+malfunction_min_duration: 3
+malfunction_max_duration: 20
+seed: 0
+regenerate_rail_on_reset: True
+regenerate_schedule_on_reset: True
\ No newline at end of file

experiments/flatland_sparse/medium_stoch_v0/tree_obs_fc_net/apex.yaml

+flatland-random-sparse-small-tree-fc-ppo:
+    run: APEX
+    env: flatland_sparse
+    stop:
+        timesteps_total: 100000000  # 1e8
+    checkpoint_freq: 10
+    checkpoint_at_end: True
+    keep_checkpoints_num: 5
+    checkpoint_score_attr: episode_reward_mean
+    config:
+        num_workers: 9
+        num_envs_per_worker: 5
+        num_gpus: 0
+
+        env_config:
+            observation: tree
+            observation_config:
+                max_depth: 2
+                shortest_path_max_depth: 30
+
+            generator: sparse_rail_generator
+            generator_config: medium_stoch_v0
+
+            wandb:
+                project: flatland
+                entity: masterscrat
+                tags: ["medium_stoch_v0", "tree_obs", "apex"] # TODO should be set programmatically
+
+        model:
+            fcnet_activation: relu
+            fcnet_hiddens: [256, 256]
+            vf_share_layers: True  # False
\ No newline at end of file