Update app.py

app.py

@@ -54,7 +54,8 @@ def generate_caption(processor, model, image, tokenizer=None):
 
 def generate_caption_coca(model, transform, image):
     im = transform(image).unsqueeze(0).to(device)
-    generated = model.generate(im, seq_len=20)
+    with torch.no_grad(), torch.cuda.amp.autocast():
+        generated = model.generate(im, seq_len=20)
     return open_clip.decode(generated[0].detach()).split("<end_of_text>")[0].replace("<start_of_text>", "")