Updated code and added LLM implementation

Q_Layered_Network/DQN_LLM.py

@@ -0,0 +1,80 @@
+import torch
+from DQN_Node_Agent import DQNAgent, QNetwork
+from transformers import GPT2LMHeadModel, GPT2Tokenizer
+
+
+class DQNGPT2LLM:
+    def __init__(self):
+        # Initialize the DQN agent
+        self.state_size = 128
+        self.action_size = 64
+        self.dqn_agent = DQNAgent(QNetwork, torch.optim.Adam,
+                                  self.state_size, self.action_size,
+                                  input_size=128, output_size=64,
+                                  gamma=0.99, min_epsilon=0.01,
+                                  epsilon_decay=0.995,
+                                  target_update_frequency=100,
+                                  epsilon=1.0)
+
+        self.training_data_path = "C:/Users/Mayra/Documents/AGI/Q_LLM/training_data/training_data.json"
+        self.training_data = DQNAgent.load_training_data(
+            self.training_data_path)
+
+        # Initialize the GPT-2 language model
+        self.tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
+        self.gpt2_model = GPT2LMHeadModel.from_pretrained("gpt2")
+
+    def train_dqn_agent(self, episodes=1000):
+        # Train DQNAgent
+        self.dqn_agent.train_dqn_agent(self.dqn_agent,
+                                       self.training_data, episodes=episodes)
+
+    def generate_response(self, query):
+        if "train" in query.lower():
+            self.train_dqn_agent()
+            return "DQN agent training completed."
+
+        elif "export" in query.lower():
+            input_example = torch.randn(1, 128)
+            self.dqn_agent.export_to_onnx(input_example.size())
+            return "DQN agent exported to ONNX."
+
+        elif "generate text" in query.lower():
+            generated_text = self.generate_text(query, max_length=100)
+            return f"Generated text from your LLM: {generated_text}"
+
+    def generate_text(self, query, max_length=40000):
+        input_ids = self.tokenizer.encode(query, return_tensors="pt")
+        attention_mask = torch.ones_like(input_ids)
+
+        output_ids = self.gpt2_model.generate(input_ids,
+                                              max_length=max_length,
+                                              num_return_sequences=1,
+                                              no_repeat_ngram_size=2,
+                                              attention_mask=attention_mask)
+
+        generated_text = self.tokenizer.decode(
+            output_ids[0], skip_special_tokens=True)
+        return generated_text
+
+
+def main():
+    dqngpt2_llm = DQNGPT2LLM()
+
+    print("BATMAN_AI CLI INTERFACE")
+
+    while True:
+        user_input = input("You: ")
+
+        if user_input.lower() == "exit":
+            print("Goodbye!")
+            break
+
+        self = dqngpt2_llm
+        query = f"{user_input}"
+        generated_text = self.generate_text(query)
+        print(f"{generated_text}")
+
+
+if __name__ == "__main__":
+    main()

Q_Layered_Network/DQN_Node_Agent.py

@@ -310,8 +310,6 @@ def train_dqn_agent(self, agent, training_data, episodes=1000):
                 next_state = torch.tensor([ord(char) for char in next_content], dtype=torch.float32)
 
 
-
-
 if __name__ == "__main__":
     action_size = 64
     state_size = 128