Spaces:
Runtime error
Runtime error
Iqra Ali
commited on
Commit
·
a394d66
1
Parent(s):
e6d304f
Update app.py
Browse files
app.py
CHANGED
@@ -1,35 +1,28 @@
|
|
1 |
|
2 |
import re
|
3 |
import gradio as gr
|
|
|
4 |
import torch
|
5 |
from transformers import DonutProcessor, VisionEncoderDecoderModel
|
6 |
-
import transformers
|
7 |
-
from PIL import Image
|
8 |
-
import random
|
9 |
-
import numpy as np
|
10 |
-
|
11 |
-
# hidde logs
|
12 |
-
transformers.logging.disable_default_handler()
|
13 |
-
|
14 |
-
|
15 |
-
# Load our model from Hugging Face
|
16 |
-
processor = DonutProcessor.from_pretrained("Iqra56/Donut_Updated")
|
17 |
-
model = VisionEncoderDecoderModel.from_pretrained("Iqra56/Donut_Updated")
|
18 |
|
19 |
-
#
|
|
|
|
|
|
|
|
|
|
|
20 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
21 |
model.to(device)
|
22 |
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
# prepare inputs
|
28 |
-
pixel_values = torch.tensor(test_sample["pixel_values"]).unsqueeze(0)
|
29 |
task_prompt = "<s>"
|
30 |
decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
|
31 |
-
|
32 |
-
#
|
33 |
outputs = model.generate(
|
34 |
pixel_values.to(device),
|
35 |
decoder_input_ids=decoder_input_ids.to(device),
|
@@ -50,18 +43,18 @@ def run_prediction(sample, model=model, processor=processor):
|
|
50 |
|
51 |
return processor.token2json(sequence)
|
52 |
|
53 |
-
description = "Gradio Demo for Donut, an instance of `VisionEncoderDecoderModel` fine-tuned on
|
54 |
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2111.15664' target='_blank'>Donut: OCR-free Document Understanding Transformer</a> | <a href='https://github.com/clovaai/donut' target='_blank'>Github Repo</a></p>"
|
55 |
|
56 |
demo = gr.Interface(
|
57 |
fn=process_document,
|
58 |
-
inputs=
|
59 |
outputs="json",
|
60 |
-
title="Demo: Donut 🍩 for
|
61 |
description=description,
|
62 |
article=article,
|
63 |
enable_queue=True,
|
64 |
-
examples=[["
|
65 |
cache_examples=False)
|
66 |
|
67 |
demo.launch()
|
|
|
1 |
|
2 |
import re
|
3 |
import gradio as gr
|
4 |
+
|
5 |
import torch
|
6 |
from transformers import DonutProcessor, VisionEncoderDecoderModel
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
8 |
+
#processor = DonutProcessor.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
|
9 |
+
#model = VisionEncoderDecoderModel.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
|
10 |
+
#processor = DonutProcessor.from_pretrained("Iqra56/ENGLISHDONUT")
|
11 |
+
#model = VisionEncoderDecoderModel.from_pretrained("Iqra56/ENGLISHDONUT")
|
12 |
+
processor = DonutProcessor.from_pretrained("Iqra56/DONUTWOKEYS")
|
13 |
+
model = VisionEncoderDecoderModel.from_pretrained("Iqra56/DONUTWOKEYS")
|
14 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
15 |
model.to(device)
|
16 |
|
17 |
+
def process_document(image):
|
18 |
+
# prepare encoder inputs
|
19 |
+
pixel_values = processor(image, return_tensors="pt").pixel_values
|
20 |
+
|
21 |
+
# prepare decoder inputs
|
|
|
22 |
task_prompt = "<s>"
|
23 |
decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
|
24 |
+
|
25 |
+
# generate answer
|
26 |
outputs = model.generate(
|
27 |
pixel_values.to(device),
|
28 |
decoder_input_ids=decoder_input_ids.to(device),
|
|
|
43 |
|
44 |
return processor.token2json(sequence)
|
45 |
|
46 |
+
description = "Gradio Demo for Donut, an instance of `VisionEncoderDecoderModel` fine-tuned on CORD (document parsing). To use it, simply upload your image and click 'submit', or click one of the examples to load them. Read more at the links below."
|
47 |
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2111.15664' target='_blank'>Donut: OCR-free Document Understanding Transformer</a> | <a href='https://github.com/clovaai/donut' target='_blank'>Github Repo</a></p>"
|
48 |
|
49 |
demo = gr.Interface(
|
50 |
fn=process_document,
|
51 |
+
inputs="image",
|
52 |
outputs="json",
|
53 |
+
title="Demo: Donut 🍩 for Document Parsing",
|
54 |
description=description,
|
55 |
article=article,
|
56 |
enable_queue=True,
|
57 |
+
examples=[[""], [""], [""]],
|
58 |
cache_examples=False)
|
59 |
|
60 |
demo.launch()
|