Upload 2 files
Browse files- ocr_gpt_app.py +24 -0
- requirements.txt +3 -0
ocr_gpt_app.py
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import easyocr
|
2 |
+
import gradio as gr
|
3 |
+
from transformers import pipeline
|
4 |
+
|
5 |
+
# Initialize OCR reader and GPT pipeline
|
6 |
+
reader = easyocr.Reader(['ch_sim', 'en'])
|
7 |
+
gpt = pipeline('text-generation', model='gpt2')
|
8 |
+
|
9 |
+
# Define the OCR function
|
10 |
+
def ocr_gpt(image):
|
11 |
+
ocr_result = reader.readtext(image)
|
12 |
+
corrected_text = gpt(ocr_result[0][1])[0]['generated_text']
|
13 |
+
gpt_opinion = gpt("What do you think about the information in the image?")[0]['generated_text']
|
14 |
+
return corrected_text, gpt_opinion
|
15 |
+
|
16 |
+
# Define Gradio UI components
|
17 |
+
image_input = gr.inputs.Image()
|
18 |
+
gpt_input = gr.inputs.Textbox(lines=1, label="GPT API Input")
|
19 |
+
ocr_output = gr.outputs.Textbox(label="OCR Result (GPT Corrected)")
|
20 |
+
gpt_opinion_output = gr.outputs.Textbox(label="GPT Opinion on Image Information")
|
21 |
+
|
22 |
+
# Create Gradio interface
|
23 |
+
iface = gr.Interface(fn=ocr_gpt, inputs=[gpt_input, image_input], outputs=[ocr_output, gpt_opinion_output])
|
24 |
+
#iface.launch(share=True)
|
requirements.txt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
easyocr
|
2 |
+
gradio
|
3 |
+
transformers
|