Commit
•
2943064
1
Parent(s):
ff9c5d7
Add image processing functionality and update dependencies
Browse files- app.py +7 -2
- book_page.jpeg +0 -0
- requirements.txt +4 -1
app.py
CHANGED
@@ -1,6 +1,8 @@
|
|
1 |
from transformers import NougatProcessor, VisionEncoderDecoderModel
|
2 |
import gradio as gr
|
3 |
import torch
|
|
|
|
|
4 |
from pdf2image import convert_from_path
|
5 |
|
6 |
# Load the model and processor
|
@@ -67,6 +69,8 @@ This is a demo for the Arabic Small Nougat model. It is an end-to-end OCR model
|
|
67 |
**Note**: The model is a prototype in my book and may not work well on all types of images and PDFs. **Check the output carefully before using it for any serious work.**
|
68 |
"""
|
69 |
|
|
|
|
|
70 |
with gr.Blocks(title="Arabic Small Nougat") as demo:
|
71 |
gr.HTML("<h1 style='text-align: center'>Arabic End-to-End Structured OCR for textbooks</h1>")
|
72 |
gr.Markdown(model_description)
|
@@ -74,10 +78,11 @@ with gr.Blocks(title="Arabic Small Nougat") as demo:
|
|
74 |
with gr.Tab("Extract Text from Image"):
|
75 |
with gr.Row():
|
76 |
with gr.Column():
|
77 |
-
|
78 |
image_submit_button = gr.Button(value="Submit", variant="primary")
|
79 |
output = gr.Markdown(label="Output Markdown", rtl=True)
|
80 |
-
image_submit_button.click(extract_text_from_image, inputs=[
|
|
|
81 |
|
82 |
with gr.Tab("Extract Text from PDF"):
|
83 |
with gr.Row():
|
|
|
1 |
from transformers import NougatProcessor, VisionEncoderDecoderModel
|
2 |
import gradio as gr
|
3 |
import torch
|
4 |
+
from PIL import Image
|
5 |
+
from pathlib import Path
|
6 |
from pdf2image import convert_from_path
|
7 |
|
8 |
# Load the model and processor
|
|
|
69 |
**Note**: The model is a prototype in my book and may not work well on all types of images and PDFs. **Check the output carefully before using it for any serious work.**
|
70 |
"""
|
71 |
|
72 |
+
example_images = [Image.open(Path(__file__).parent / "book_page.jpeg")]
|
73 |
+
|
74 |
with gr.Blocks(title="Arabic Small Nougat") as demo:
|
75 |
gr.HTML("<h1 style='text-align: center'>Arabic End-to-End Structured OCR for textbooks</h1>")
|
76 |
gr.Markdown(model_description)
|
|
|
78 |
with gr.Tab("Extract Text from Image"):
|
79 |
with gr.Row():
|
80 |
with gr.Column():
|
81 |
+
input_image = gr.Image(label="Input Image", type="pil")
|
82 |
image_submit_button = gr.Button(value="Submit", variant="primary")
|
83 |
output = gr.Markdown(label="Output Markdown", rtl=True)
|
84 |
+
image_submit_button.click(extract_text_from_image, inputs=[input_image], outputs=output)
|
85 |
+
gr.Examples(example_images, [input_image], output, extract_text_from_image, cache_examples=True)
|
86 |
|
87 |
with gr.Tab("Extract Text from PDF"):
|
88 |
with gr.Row():
|
book_page.jpeg
ADDED
![]() |
requirements.txt
CHANGED
@@ -2,4 +2,7 @@ pdf2image
|
|
2 |
torch
|
3 |
transformers
|
4 |
gradio
|
5 |
-
python-Levenshtein
|
|
|
|
|
|
|
|
2 |
torch
|
3 |
transformers
|
4 |
gradio
|
5 |
+
python-Levenshtein
|
6 |
+
pillow
|
7 |
+
pathlib
|
8 |
+
nltk
|