Update app.py

app.py

@@ -52,11 +52,21 @@ def image_preprocess(image):
 
 def generate_caption(processor, model, device, image):
     inputs =  image_processor (image, return_tensors='pt').to(device)
-    out = model.generate(**inputs, max_new_tokens=20)
-    caption = processor.decode(out[0], skip_special_tokens=True)
-    #caption="im here "
+    model.eval()
+    # Generate caption
+    with torch.no_grad():
+      output = model.generate(
+          pixel_values=inputs ,
+          max_length=1000,  # Adjust the maximum length of the generated caption as needed
+          num_beams=4,    # Adjust the number of beams for beam search decoding
+          early_stopping=True  # Enable early stopping to stop generation when all beams finished
+      )
+
+      # Decode the generated caption
+      caption = tokenizer.decode(output[0], skip_special_tokens=True)
     return caption
 
+
 def main():
     set_page_config()
     st.header("Caption an Image :camera:")