Update app.py

app.py

@@ -29,8 +29,8 @@ outputs = [
 examples = [['img_1.jpg','img_2.jpg']]
 
 def get_output_senten(img):
-  pixel_values = vit_feature_extractor(images=img, return_tensors="pt").pixel_values.to('cuda')
-  encoder_outputs = model.generate(pixel_values.to('cuda'),num_beams=7)
+  pixel_values = vit_feature_extractor(images=img, return_tensors="pt").pixel_values.to('cpu')
+  encoder_outputs = model.generate(pixel_values.to('cpu'),num_beams=7)
   generated_sentences = tokenizer.batch_decode(encoder_outputs)
   senten = generated_sentences[0][generated_sentences[0][2:].index('>')+1:]