Update app.py

app.py

@@ -147,8 +147,13 @@ except Exception as e:
 def generate_response(model, tokenizer, instruction, max_new_tokens=2048):
     """Generate a response from the model based on an instruction."""
     try:
-        # Format input for the model
-        inputs = tokenizer.encode(instruction, return_tensors="pt").to(model.device)
+        # Encode input with truncation
+        inputs = tokenizer.encode(
+            instruction,
+            return_tensors="pt",
+            truncation=True,
+            max_length=tokenizer.model_max_length
+        ).to(model.device)
         
         # Generate response
         outputs = model.generate(
@@ -159,13 +164,19 @@ def generate_response(model, tokenizer, instruction, max_new_tokens=2048):
             do_sample=True,
         )
 
-        # Decode and return the output
+        # Decode and strip input prompt from response
         response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-        return response
+        generated_text = response[len(instruction):].strip()
+        
+        print(f"Instruction: {instruction}")  # Debugging line
+        print(f"Generated Response: {generated_text}")  # Debugging line
+
+        return generated_text
 
     except Exception as e:
         raise ValueError(f"Error generating response: {e}")
 
+
 @app.post("/generate")
 async def generate_text(input: ModelInput):
     try: