Updates

Shared.py

@@ -12,7 +12,7 @@ class Shared:
         gamma=0.99,
         epsilon=0.1,
         run_name=None,
-        frozenlake_size=8,
+        frozenlake_size=16,
         **kwargs,
     ):
         print("=" * 80)

run.py

@@ -108,12 +108,6 @@ def main():
         default=None,
         help="WandB project name for logging. If not provided, no logging is done. (default: None)",
     )
-    parser.add_argument(
-        "--wandb_group",
-        type=str,
-        default="monte-carlo",
-        help="WandB group name for logging. (default: monte-carlo)",
-    )
     parser.add_argument(
         "--wandb_job_type",
         type=str,
@@ -144,7 +138,7 @@ def main():
                 wandb.init(
                     project=args.wandb_project,
                     name=run_name,
-                    group=args.wandb_group,
+                    group=args.agent,
                     job_type=args.wandb_job_type,
                     config=dict(args._get_kwargs()),
                 )

run_tests_MC_CliffWalking-v0.py

@@ -19,7 +19,7 @@ vals_gamma = [1.0, 0.98, 0.96, 0.94]
 
 def run_test(args):
     os.system(
-        f"python3 MonteCarloAgent.py --train --n_train_episodes {n_train_episodes} --max_steps {max_steps} --env {env} --gamma {args[0]} --epsilon {args[1]} --update_type {args[2]} --wandb_project {wandb_project} --wandb_run_name_suffix {args[3]} --no_save"
+        f"python3 MonteCarloAgent.py --agent MCAgent --train --n_train_episodes {n_train_episodes} --max_steps {max_steps} --env {env} --gamma {args[0]} --epsilon {args[1]} --update_type {args[2]} --wandb_project {wandb_project} --wandb_run_name_suffix {args[3]} --no_save"
     )
 
 

run_tests_MC_FrozenLake-v1.py

@@ -17,7 +17,7 @@ vals_gamma = [1.0, 0.98, 0.96, 0.94]
 
 def run_test(args):
     os.system(
-        f"python3 MonteCarloAgent.py --train --n_train_episodes {n_train_episodes} --max_steps {max_steps} --env {env} --gamma {args[0]} --epsilon {args[1]} --update_type {args[2]} --wandb_project {wandb_project} --wandb_run_name_suffix {args[3]} --no_save"
+        f"python3 MonteCarloAgent.py --agent MCAgent --train --n_train_episodes {n_train_episodes} --max_steps {max_steps} --env {env} --gamma {args[0]} --epsilon {args[1]} --update_type {args[2]} --wandb_project {wandb_project} --wandb_run_name_suffix {args[3]} --no_save"
     )