configs updated and added for ai4realnet_small env

Author: Mleyliabadi

config.ini

@@ -21,3 +21,10 @@ eval_kwargs = {
     "return_episode_rewards": True,
     }
 
+[TRAIN]
+train_kwargs = {
+    "model_name": "PPO_SB3",
+    "train_steps": 1000000,
+    "save_freq": 100000,    
+    "save_path": "trained_models/ai4realnet_small/custom_agent",
+    }

config_ai4realnet_small.ini

@@ -0,0 +1,53 @@
+[DEFAULT]
+kpi_kwargs = {
+    "env_name": "ai4realnet_small",
+    "acceptance_threshold": 200,
+    "finetune_budget": 100000,
+    "save_path": "results/ai4realnet_small"
+    }
+
+train_kwargs = {
+    "model_name": "PPO_SB3",
+    "train_steps": 1000,
+    "save_freq": 5000,    
+    "load_path": "trained_models/ai4realnet_small/PPO_SB3/PPO_SB3.zip",
+    "save_path": "trained_models/ai4realnet_small/PPO_SB3_FINETUNE",
+    }
+    
+eval_kwargs = {
+    "n_eval_episodes": 50,
+    "render": False,
+    "deterministic": True,
+    "return_episode_rewards": True,
+    }
+
+[EXPERT_AGENT]
+kpi_kwargs = {
+    "env_name": "ai4realnet_small",
+    "acceptance_threshold": 200,
+    "finetune_budget": 100000,
+    "save_path": "results/ai4realnet_small/expert_agent"
+    }
+
+train_kwargs = {
+    "train_steps": 5000,
+    "n_eval_episodes": 2,
+    "save_freq": 5000,    
+    "load_path": "trained_models/ai4realnet_small/expert_agent/EXPERT_AGENT.zip",
+    "save_path": "trained_models/ai4realnet_small/expert_agent/EXPERT_AGENT_FINETUNE",
+    "load_path_fine_tune": "trained_models/ai4realnet_small/expert_agent/EXPERT_AGENT_FINETUNE/PPO_SB3.zip"
+    }
+
+eval_kwargs = {
+    "nb_episode": 2,
+    "verbose": True,
+    "logs_path": "trained_models/ai4realnet_small/expert_agent/eval_logs",
+    "nb_process": 1,
+    }
+; eval_kwargs = {
+;     "n_eval_episodes": 50,
+;     "render": False,
+;     "deterministic": True,
+;     "return_episode_rewards": True,
+;     "logs_path": "trained_models/ai4realnet_small/expert_agent"
+;     }