Update checkpoints, rollout scripts, add test_result script

baselines/test_results.py

+import numpy as np
+import pandas as pd
+import os
+import json
+
+# The file all_eval_runs are generated by the wandb_data.py
+df_eval = pd.read_csv('all_eval_runs.csv')
+df_test_results = df_eval[["run","group"]].drop_duplicates()
+
+all_runs = df_test_results["run"].to_list()
+
+colnames = ["run","percentage_complete_mean","normalized_reward_mean"]
+df_test_metrics = pd.DataFrame(columns= colnames)
+for cur_run in all_runs:
+    result_file = "checkpoints/"+ cur_run + "/test_outcome.json"
+    if os.path.isfile(result_file):
+        with open(result_file) as f:
+            data = json.load(f)
+        df_test_metrics = df_test_metrics.append({colnames[0]:cur_run,colnames[1]:data.get(colnames[1]),colnames[2]:data.get(colnames[2])},ignore_index = True)
+        
+df_test = pd.merge(df_test_metrics,df_test_results,how='left')
+
+df_all_final_results = df_test.groupby("group").aggregate([np.mean,np.std]).reset_index()
+
+df_all_final_results.to_csv('test_results_group.csv',index=False)

rollout.py

@@ -13,6 +13,7 @@ import gym
 import numpy as np
 import ray
 from ray.rllib.agents.registry import get_agent_class
+from ray.tune.registry import get_trainable_cls
 from ray.rllib.env import MultiAgentEnv
 from ray.rllib.env.base_env import _DUMMY_AGENT_ID
 # from ray.rllib.evaluation.episode import _flatten_action # ray 0.8.4
@@ -20,7 +21,6 @@ from ray.rllib.policy.sample_batch import DEFAULT_POLICY_ID
 from ray.rllib.utils.space_utils import flatten_to_single_ndarray # ray 0.8.5
 from ray.tune.utils import merge_dicts
 
-from algorithms.imitation_agent.imitation_trainer import ImitationAgent
 from utils.loader import load_envs, load_models, load_algorithms
 
 logger = logging.getLogger(__name__)
@@ -44,6 +44,23 @@ load_models(os.getcwd())  # Load models
 from algorithms import CUSTOM_ALGORITHMS
 load_algorithms(CUSTOM_ALGORITHMS)  # Load algorithms
 
+from collections.abc import Mapping
+from copy import deepcopy
+
+def val_replace(mapping):
+    obj = deepcopy(mapping)
+    if isinstance(mapping, Mapping):
+        for key, val in mapping.items():
+            obj[key] = val_replace(val)
+    else:
+        if mapping == "False":
+            return False
+        if mapping == "True":
+            return True
+        else:
+            return mapping
+    return obj
+
 
 class RolloutSaver:
     """Utility class for storing rollouts.
@@ -233,6 +250,10 @@ def create_parser(parser_creator=None):
         help="Write progress to a temporary file (updated "
              "after each episode). An output filename must be set using --out; "
              "the progress file will live in the same folder.")
+    parser.add_argument(
+        "--eager",
+        action="store_true",
+        help="Whether to attempt to enable TF eager execution.")
     return parser
 
 
@@ -253,26 +274,22 @@ def run(args, parser):
             config = pickle.load(f)
     if "num_workers" in config:
         config["num_workers"] = min(2, config["num_workers"])
-    config = merge_dicts(config, args.config)
+
+    updated_config = val_replace(args.config)
+    config = merge_dicts(config, updated_config)
     if not args.env:
         if not config.get("env"):
             parser.error("the following arguments are required: --env")
         args.env = config.get("env")
 
     ray.init()
-    try:
-        cls = get_agent_class(args.run)
-    except:
-        cls = ImitationAgent # CUSTOM_ALGORITHMS[args.run]
+    
+    if args.eager:
         from tensorflow.python.framework.ops import enable_eager_execution
         enable_eager_execution()
-        print("========================")
-        print(config)
-        print("========================")
-        print(config.get("model").get("vf_share_layers"))
-        print("========================")
-        config['model']['vf_share_layers'] = False
-        print("========================")
+    
+    cls = get_trainable_cls(args.run)
+    
     agent = cls(env=args.env, config=config)
     agent.restore(args.checkpoint)
     num_steps = int(args.steps)