Add tqdm bar to rollout.py

evaluator/rollout.py

@@ -13,6 +13,8 @@ from config import LocalEvaluationConfig
 import random
 import copy
 
+from tqdm import tqdm
+
 class ProxyEnv(Env):
     
    def step(self, actions, preprocess=set(), omitDead=True):
@@ -75,6 +77,7 @@ def assign_agents(player_agent, opponent_agents):
 
 def run_episode(player_index, agents, N_TIME_STEPS):
     config = projekt.config.CompetitionRound1()
+    config.RENDER = True
     env = ProxyEnv(config)
     n_steps = 0
     neural_agents = set()
@@ -92,9 +95,9 @@ def run_episode(player_index, agents, N_TIME_STEPS):
     for entid in entids:
         actions[entid] = agents[entid_agent_map[entid]].register_reset(obs[entid])
     alive_agents = list(obs.keys())
-
+    pbar = tqdm(total = N_TIME_STEPS)
     while len(obs.keys()) > 0 and n_steps < N_TIME_STEPS:
-
+        pbar.update(n=1)
         for entid in actions:
             if entid not in neural_agents:
                 realm = env.realm
@@ -157,7 +160,7 @@ if __name__ == "__main__":
     player_agent, opponent_agents = load_agents(LocalEvaluationConfig)
     player_agent = player_agent[0]
     N_EPISODES = 2
-    N_TIME_STEPS = 10
+    N_TIME_STEPS = 1024
     for episode in range(N_EPISODES):
         agents, player_index = assign_agents(player_agent, opponent_agents)
         statistics = run_episode(player_index, agents, N_TIME_STEPS)