Update app.py

app.py

@@ -73,8 +73,8 @@ def load_model():
         raise
 
 # Load the model and tokenizer locally
-model, tokenizer = load_model()
-#model = HfApiModel(model_id='mistralai/Mistral-7B-Instruct-v0.2', max_tokens=512)
+#model, tokenizer = load_model()
+model = HfApiModel(model_id='mistralai/Mistral-7B-Instruct-v0.2', max_tokens=512)
 
 
 class BasicAgent:
@@ -89,22 +89,11 @@ class BasicAgent:
             planning_interval=3,
             
         )
-        
-    def __call__(self, question: str) -> str:
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
-
-        encoded_input = tokenizer(question, return_tensors="pt", padding=True, truncation=True, max_length=512)
 
-        # Generate text using the model
-        output = model.generate(
-            encoded_input['input_ids'],
-            attention_mask=encoded_input['attention_mask'],  # Pass attention mask
-            max_length=100,
-            pad_token_id=tokenizer.pad_token_id,
 
-        )
-
-        answer = tokenizer.decode(output[0], skip_special_tokens=True)
+    def __call__(self, question: str) -> str:
+        print(f"Agent received question (first 50 chars): {question[:50]}...")
+        answer = self.agent.run(question)
         return answer