code clean up

README.md

@@ -1,5 +1,5 @@
 ---
-title: Donut Cn Invoice
+title: Donut for Chinese Invoice
 emoji: 🌍
 colorFrom: red
 colorTo: blue

app.py

@@ -5,11 +5,8 @@ import gradio as gr
 import torch
 from PIL import Image
 
-print("test1")
 processor = DonutProcessor.from_pretrained("ewfian/donut_cn_invoice")
-print("test2")
 model = VisionEncoderDecoderModel.from_pretrained("ewfian/donut_cn_invoice")
-print("test3")
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 model.to(device)
@@ -17,17 +14,10 @@ model.to(device)
 task_prompt = "<s_totalAmountInWords>"
 decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
 
-print("test")
-print(decoder_input_ids.shape)
-
 def process_document(image):
-    print("test2")
 
     pixel_values = processor(image, return_tensors="pt").pixel_values
 
-    print(pixel_values.shape)
-    print(pixel_values)
-
     outputs = model.generate(
         pixel_values.to(device),
         decoder_input_ids=decoder_input_ids.to(device),
@@ -44,9 +34,6 @@ def process_document(image):
     sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()  # remove first task start token
     return processor.token2json(sequence)
 
-# t = process_document(test_sample)
-# print(t)
-
 demo = gr.Interface(
     fn=process_document,
     inputs="image",

requirements.txt

@@ -1,7 +1,6 @@
 sentencepiece==0.1.99
 transformers==4.35.0
 protobuf==4.23.4
-gradio<4.0
 -f https://download.pytorch.org/whl/torch_stable.html 
 torch==2.1.0+cpu 
 -f https://download.pytorch.org/whl/torch_stable.html