Spaces:
Sleeping
Sleeping
Adding the necessary points
Browse files- README.md +5 -5
- app.py +99 -0
- requirements.txt +5 -0
- test_images/test_images/test_0.jpg +0 -0
- test_images/test_images/test_1.jpg +0 -0
- test_images/test_images/test_2.jpg +0 -0
- test_images/test_images/test_3.jpg +0 -0
- test_images/test_images/test_4.jpg +0 -0
- test_images/test_images/test_5.jpg +0 -0
- test_images/test_images/test_6.jpg +0 -0
- test_images/test_images/test_7.jpg +0 -0
- test_images/test_images/test_8.jpg +0 -0
- test_images/test_images/test_9.jpg +0 -0
README.md
CHANGED
@@ -1,10 +1,10 @@
|
|
1 |
---
|
2 |
-
title:
|
3 |
-
emoji:
|
4 |
-
colorFrom:
|
5 |
-
colorTo:
|
6 |
sdk: gradio
|
7 |
-
sdk_version: 4.
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
|
|
1 |
---
|
2 |
+
title: Donut Prototype
|
3 |
+
emoji: π
|
4 |
+
colorFrom: red
|
5 |
+
colorTo: red
|
6 |
sdk: gradio
|
7 |
+
sdk_version: 4.24.0
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
app.py
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# -*- coding: utf-8 -*-
|
2 |
+
"""Demo.ipynb
|
3 |
+
|
4 |
+
Automatically generated by Colaboratory.
|
5 |
+
|
6 |
+
Original file is located at
|
7 |
+
https://colab.research.google.com/drive/1Icb8zeoaudyTDOKM1QySNay1cXzltRAp
|
8 |
+
"""
|
9 |
+
|
10 |
+
import gradio as gr
|
11 |
+
from PIL import Image
|
12 |
+
import re
|
13 |
+
|
14 |
+
import torch
|
15 |
+
import torch.nn as nn
|
16 |
+
from warnings import simplefilter
|
17 |
+
|
18 |
+
simplefilter('ignore')
|
19 |
+
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
20 |
+
|
21 |
+
# Seting up the model
|
22 |
+
from transformers import DonutProcessor, VisionEncoderDecoderModel
|
23 |
+
|
24 |
+
print('Loading the base model ....')
|
25 |
+
base_model = VisionEncoderDecoderModel.from_pretrained('Edgar404/donut-shivi-recognition')
|
26 |
+
base_processor = DonutProcessor.from_pretrained('Edgar404/donut-shivi-recognition')
|
27 |
+
print('Loading complete')
|
28 |
+
|
29 |
+
print('Loading the optimized model ....')
|
30 |
+
optimized_model = VisionEncoderDecoderModel.from_pretrained('Edgar404/donut-shivi-cheques_KD_320', torch_dtype = torch.bfloat16 )
|
31 |
+
optimized_processor = DonutProcessor.from_pretrained('Edgar404/donut-shivi-cheques_KD_320')
|
32 |
+
print('Loading complete')
|
33 |
+
|
34 |
+
# setting
|
35 |
+
|
36 |
+
|
37 |
+
def process_image(image , mode = 'optimized' ):
|
38 |
+
""" Function that takes an image and perform an OCR using the model DonUT via the task document
|
39 |
+
parsing
|
40 |
+
|
41 |
+
parameters
|
42 |
+
__________
|
43 |
+
image : a machine readable image of class PIL or numpy"""
|
44 |
+
|
45 |
+
model = optimized_model if mode == 'optimized' else base_model
|
46 |
+
processor = optimized_processor if mode == 'optimized' else base_processor
|
47 |
+
d_type = torch.bfloat16 if (mode == 'optimized' & device =='cuda') else torch.float32
|
48 |
+
|
49 |
+
model.to(device)
|
50 |
+
model.eval()
|
51 |
+
|
52 |
+
|
53 |
+
task_prompt = "<s_cord-v2>"
|
54 |
+
decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
|
55 |
+
|
56 |
+
pixel_values = processor(image, return_tensors="pt").pixel_values
|
57 |
+
|
58 |
+
outputs = model.generate(
|
59 |
+
pixel_values.to(device , dtype = d_type),
|
60 |
+
decoder_input_ids=decoder_input_ids.to(device),
|
61 |
+
max_length=model.decoder.config.max_position_embeddings,
|
62 |
+
pad_token_id=processor.tokenizer.pad_token_id,
|
63 |
+
eos_token_id=processor.tokenizer.eos_token_id,
|
64 |
+
use_cache=True,
|
65 |
+
bad_words_ids=[[processor.tokenizer.unk_token_id]],
|
66 |
+
return_dict_in_generate=True,
|
67 |
+
)
|
68 |
+
|
69 |
+
sequence = processor.batch_decode(outputs.sequences)[0]
|
70 |
+
sequence = sequence.replace(processor.tokenizer.eos_token, "").replace(processor.tokenizer.pad_token, "")
|
71 |
+
sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()
|
72 |
+
output = processor.token2json(sequence)
|
73 |
+
|
74 |
+
return output
|
75 |
+
|
76 |
+
|
77 |
+
def image_classifier(image , mode):
|
78 |
+
return process_image(image , mode)
|
79 |
+
|
80 |
+
|
81 |
+
|
82 |
+
examples_list = [['./test_images/test_0.jpg' ,"base"] ,
|
83 |
+
['./test_images/test_1.jpg','base'],
|
84 |
+
['./test_images/test_2.jpg' ,"base"],
|
85 |
+
['./test_images/test_3.jpg','base'],
|
86 |
+
['./test_images/test_4.jpg','base'],
|
87 |
+
['./test_images/test_5.jpg' ,"base"],
|
88 |
+
['./test_images/test_6.jpg' ,"base"],
|
89 |
+
['./test_images/test_7.jpg','base'],
|
90 |
+
['./test_images/test_8.jpg','base'],
|
91 |
+
['./test_images/test_9.jpg','base']
|
92 |
+
]
|
93 |
+
|
94 |
+
demo = gr.Interface(fn=image_classifier, inputs=["image",
|
95 |
+
gr.Radio(["base" , "optimized"], label="mode")],
|
96 |
+
outputs="text",
|
97 |
+
examples = examples_list )
|
98 |
+
|
99 |
+
demo.launch(share = True , debug = True)
|
requirements.txt
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
transformers
|
2 |
+
torch
|
3 |
+
pillow
|
4 |
+
gradio
|
5 |
+
peft
|
test_images/test_images/test_0.jpg
ADDED
test_images/test_images/test_1.jpg
ADDED
test_images/test_images/test_2.jpg
ADDED
test_images/test_images/test_3.jpg
ADDED
test_images/test_images/test_4.jpg
ADDED
test_images/test_images/test_5.jpg
ADDED
test_images/test_images/test_6.jpg
ADDED
test_images/test_images/test_7.jpg
ADDED
test_images/test_images/test_8.jpg
ADDED
test_images/test_images/test_9.jpg
ADDED