Update app.py

app.py

@@ -62,28 +62,30 @@ if file_name is not None:
     processor_ext = DonutProcessor.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
     model_ext = VisionEncoderDecoderModel.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
 
+    device = "cuda" if torch.cuda.is_available() else "cpu"
     model_ext.to(device)
-   
+
     # prepare decoder inputs
     task_prompt = "<s_cord-v2>"
-    decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
-    
+    decoder_input_ids = processor_ext.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
+
     pixel_values = processor_ext(image, return_tensors="pt").pixel_values
 
     outputs = model_ext.generate(
         pixel_values.to(device),
         decoder_input_ids=decoder_input_ids.to(device),
-        max_length=model.decoder.config.max_position_embeddings,
+        max_length=model_ext.decoder.config.max_position_embeddings,
         pad_token_id=processor.tokenizer.pad_token_id,
         eos_token_id=processor.tokenizer.eos_token_id,
         use_cache=True,
-        bad_words_ids=[[processor.tokenizer.unk_token_id]],
+        bad_words_ids=[[processor_ext.tokenizer.unk_token_id]],
         return_dict_in_generate=True,
     )
 
     sequence = processor_ext.batch_decode(outputs.sequences)[0]
     sequence = sequence.replace(processor_ext.tokenizer.eos_token, "").replace(processor_ext.tokenizer.pad_token, "")
     sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()  # remove first task start token
+    print(processor_ext.token2json(sequence))
     col3.header("Features")
     col3.subheader(processor_ext.token2json(sequence))