Spaces:
Runtime error
Runtime error
choiyk0103
commited on
Commit
•
76834e5
1
Parent(s):
08308ec
Delete app_py
Browse files
app_py
DELETED
@@ -1,42 +0,0 @@
|
|
1 |
-
import gradio as gr
|
2 |
-
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
3 |
-
import requests
|
4 |
-
from PIL import Image
|
5 |
-
|
6 |
-
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
|
7 |
-
model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-handwritten")
|
8 |
-
|
9 |
-
# load image examples
|
10 |
-
urls = ['https://fki.tic.heia-fr.ch/static/img/a01-122-02.jpg', 'https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSoolxi9yWGAT5SLZShv8vVd0bz47UWRzQC19fDTeE8GmGv_Rn-PCF1pP1rrUx8kOjA4gg&usqp=CAU',
|
11 |
-
'https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRNYtTuSBpZPV_nkBYPMFwVVD9asZOPgHww4epu9EqWgDmXW--sE2o8og40ZfDGo87j5w&usqp=CAU']
|
12 |
-
for idx, url in enumerate(urls):
|
13 |
-
image = Image.open(requests.get(url, stream=True).raw)
|
14 |
-
image.save(f"image_{idx}.png")
|
15 |
-
|
16 |
-
def process_image(image):
|
17 |
-
# prepare image
|
18 |
-
pixel_values = processor(image, return_tensors="pt").pixel_values
|
19 |
-
|
20 |
-
# generate (no beam search)
|
21 |
-
generated_ids = model.generate(pixel_values)
|
22 |
-
|
23 |
-
# decode
|
24 |
-
generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
25 |
-
|
26 |
-
return generated_text
|
27 |
-
|
28 |
-
title = "Interactive demo: TrOCR"
|
29 |
-
description = "Demo for Microsoft's TrOCR, an encoder-decoder model consisting of an image Transformer encoder and a text Transformer decoder for state-of-the-art optical character recognition (OCR) on single-text line images. This particular model is fine-tuned on IAM, a dataset of annotated handwritten images. To use it, simply upload a (single-text line) image or use one of the example images below and click 'submit'. Results will show up in a few seconds."
|
30 |
-
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2109.10282'>TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models</a> | <a href='https://github.com/microsoft/unilm/tree/master/trocr'>Github Repo</a></p>"
|
31 |
-
examples =[["image_0.png"], ["image_1.png"], ["image_2.png"]]
|
32 |
-
|
33 |
-
#css = """.output_image, .input_image {height: 600px !important}"""
|
34 |
-
|
35 |
-
iface = gr.Interface(fn=process_image,
|
36 |
-
inputs=gr.inputs.Image(type="pil"),
|
37 |
-
outputs=gr.outputs.Textbox(),
|
38 |
-
title=title,
|
39 |
-
description=description,
|
40 |
-
article=article,
|
41 |
-
examples=examples)
|
42 |
-
iface.launch(debug=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|