Update app.py

app.py

@@ -38,16 +38,16 @@ generation_config = GenerationConfig(
 
 if instruction:
     with torch.no_grad():
-    outputs = model.generate(
-        input_ids=input_ids,
-        attention_mask=torch.ones_like(input_ids),
-        generation_config=generation_config,
-        return_dict_in_generate=True, 
-        output_scores=True
-    )
+        outputs = model.generate(
+            input_ids=input_ids,
+            attention_mask=torch.ones_like(input_ids),
+            generation_config=generation_config,
+            return_dict_in_generate=True, 
+            output_scores=True
+        )
     
     output_text = tokenizer.decode(
-        outputs['sequences'][0].cuda(), 
+        outputs['sequences'][0],#.cuda(), 
         skip_special_tokens=False
     ).strip()