Handling Signal

app.py

@@ -30,17 +30,19 @@ def main():
 
     if start_button:
         agent = perform_training(jammer_type, channel_switching_cost)
+        st.subheader("Generating Insights of the DRL-Training")
+        model_name = "tiiuae/falcon-7b-instruct"
+        model = TFAutoModelForCausalLM.from_pretrained(model_name)
+        tokenizer = AutoTokenizer.from_pretrained(model_name)
+        pipeline = transformers.pipeline("text-generation", model=model, tokenizer=tokenizer, max_length=100,
+                                         temperature=0.7)
+        text = pipeline("Discuss this topic: Integrating LLMs to DRL-based anti-jamming.")
+        st.write(text)
         test(agent, jammer_type, channel_switching_cost)
 
+
 def perform_training(jammer_type, channel_switching_cost):
     agent = train(jammer_type, channel_switching_cost)
-    st.subheader("Generating Insights of the DRL-Training")
-    model_name = "tiiuae/falcon-7b-instruct"
-    model = TFAutoModelForCausalLM.from_pretrained(model_name)
-    tokenizer = TFAutoTokenizer.from_pretrained(model_name)
-    pipeline = transformers.pipeline("text-generation", model=model, tokenizer=tokenizer, max_length=100, temperature=0.7)
-    text = pipeline("Discuss this topic: Integrating LLMs to DRL-based anti-jamming.")
-    st.write(text)
     return agent
 
 

trainer.py

@@ -21,7 +21,7 @@ def train(jammer_type, channel_switching_cost):
     s_size = ob_space.shape[0]
     a_size = ac_space.n
     max_env_steps = 100
-    TRAIN_Episodes = 10
+    TRAIN_Episodes = 5
     env._max_episode_steps = max_env_steps
 
     epsilon = 1.0