Result format

app.py

@@ -132,22 +132,19 @@ async def predict(q: str):
         )
 
         # Get and decode the output
-        print(generated_ids)
-        output_ids = generated_ids[0][-1]
-        logger.debug(output_ids)
-        #[len(model_inputs.input_ids[0]) :]
+        output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :]
         result = tokenizer.decode(output_ids, skip_special_tokens=True)
 
         # Checkpoint
         processing_time = time.time() - start_time
 
         return ModelResponse(
-            text=result['label'],
-            confidence=result['score'],
+            text=result, #['label'],
+            confidence=0, #result['score'],
             processing_time=processing_time
         )
 
-    except Exception as e:
+    except HTTPException as e:
         logger.error(f"Evaluation error: {e}")
         raise HTTPException(status_code=500, detail="Evaluation failed")