Update app.py

app.py

@@ -22,7 +22,6 @@ model = LlavaNextForConditionalGeneration.from_pretrained(
     low_cpu_mem_usage=True,
 )
 model = PeftModel.from_pretrained(model, finetune_repo)
-# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 model.to("cuda:0")
 
 
@@ -48,12 +47,6 @@ def predict(image, input_text):
         time.sleep(0.04)
         yield generated_text_without_prompt
 
-
-    # prompt_length = inputs['input_ids'].shape[1]
-    # generate_ids = model.generate(**inputs, max_new_tokens=512)
-    # output_text = processor.batch_decode(generate_ids[:, prompt_length:], skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
-    # return output_text
-
    
 image = gr.components.Image(type="pil")
 input_prompt = gr.components.Textbox(label="Input Prompt")