Spaces:
Build error
Build error
linjieccc
commited on
Commit
·
c757801
1
Parent(s):
128bcf9
update app.py
Browse files- app.py +24 -26
- header.html +3 -7
- paddlenlp-preview.jpeg +0 -3
app.py
CHANGED
|
@@ -112,40 +112,40 @@ class PdfReader(object):
|
|
| 112 |
|
| 113 |
examples = [
|
| 114 |
[
|
| 115 |
-
"
|
| 116 |
-
"
|
| 117 |
],
|
| 118 |
[
|
| 119 |
-
"
|
| 120 |
-
"
|
| 121 |
],
|
| 122 |
[
|
| 123 |
-
"
|
| 124 |
-
"
|
| 125 |
],
|
| 126 |
[
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
],
|
| 130 |
[
|
| 131 |
-
"
|
| 132 |
-
"
|
| 133 |
],
|
| 134 |
[
|
| 135 |
-
"
|
| 136 |
-
"
|
| 137 |
],
|
| 138 |
[
|
| 139 |
-
"
|
| 140 |
-
"
|
| 141 |
],
|
| 142 |
[
|
| 143 |
-
"
|
| 144 |
-
"
|
| 145 |
],
|
| 146 |
[
|
| 147 |
-
"
|
| 148 |
-
"
|
| 149 |
],
|
| 150 |
]
|
| 151 |
|
|
@@ -250,10 +250,10 @@ def process_prompt(prompt, document, lang="ch"):
|
|
| 250 |
)
|
| 251 |
|
| 252 |
|
| 253 |
-
def load_example_document(img, prompt):
|
| 254 |
if img is not None:
|
| 255 |
document = prompt_files[prompt]
|
| 256 |
-
preview, answer, answer_text = process_prompt(prompt, document)
|
| 257 |
return document, prompt, preview, gr.update(visible=True), answer, answer_text
|
| 258 |
else:
|
| 259 |
return None, None, None, gr.update(visible=False), None, None
|
|
@@ -355,9 +355,8 @@ gradio-app h2, .gradio-app h2 {
|
|
| 355 |
with gr.Blocks(css=CSS) as demo:
|
| 356 |
gr.HTML(read_content("header.html"))
|
| 357 |
gr.Markdown(
|
| 358 |
-
f" ⚡DocPrompt⚡ is a Document Prompt Engine
|
| 359 |
-
f"
|
| 360 |
-
f" click one of the examples to load them."
|
| 361 |
)
|
| 362 |
|
| 363 |
document = gr.Variable()
|
|
@@ -463,7 +462,7 @@ with gr.Blocks(css=CSS) as demo:
|
|
| 463 |
|
| 464 |
prompt.submit(
|
| 465 |
fn=process_prompt,
|
| 466 |
-
inputs=[prompt, document],
|
| 467 |
outputs=[image, output, output_text],
|
| 468 |
)
|
| 469 |
|
|
@@ -481,11 +480,10 @@ with gr.Blocks(css=CSS) as demo:
|
|
| 481 |
|
| 482 |
example_image.change(
|
| 483 |
fn=load_example_document,
|
| 484 |
-
inputs=[example_image, example_prompt],
|
| 485 |
outputs=[document, prompt, image, img_clear_button, output, output_text],
|
| 486 |
)
|
| 487 |
|
| 488 |
-
gr.Image('./paddlenlp-preview.jpeg')
|
| 489 |
gr.Markdown("[](https://github.com/PaddlePaddle/PaddleNLP)")
|
| 490 |
gr.HTML(read_content("footer.html"))
|
| 491 |
|
|
|
|
| 112 |
|
| 113 |
examples = [
|
| 114 |
[
|
| 115 |
+
"budget_form.png",
|
| 116 |
+
"What is the total actual and/or obligated expenses of ECG Center?"
|
| 117 |
],
|
| 118 |
[
|
| 119 |
+
"medical_bill_2.png",
|
| 120 |
+
"患者さんは何でお金を払いますか。"
|
| 121 |
],
|
| 122 |
[
|
| 123 |
+
"receipt.png",
|
| 124 |
+
"เบอร์โทรร้านอะไรคะ"
|
| 125 |
],
|
| 126 |
[
|
| 127 |
+
"poster.png",
|
| 128 |
+
"Which gift idea needs a printer?"
|
| 129 |
],
|
| 130 |
[
|
| 131 |
+
"resume.png",
|
| 132 |
+
"五百丁本次想要担任的是什么职位?",
|
| 133 |
],
|
| 134 |
[
|
| 135 |
+
"custom_declaration_form.png",
|
| 136 |
+
"在哪个口岸进口?"
|
| 137 |
],
|
| 138 |
[
|
| 139 |
+
"invoice.jpg",
|
| 140 |
+
"发票号码是多少?",
|
| 141 |
],
|
| 142 |
[
|
| 143 |
+
"medical_bill_1.png",
|
| 144 |
+
"票据的具体名称是什么?"
|
| 145 |
],
|
| 146 |
[
|
| 147 |
+
"website_design_guide.jpeg",
|
| 148 |
+
"Which quality component has the icon of a pen in it?"
|
| 149 |
],
|
| 150 |
]
|
| 151 |
|
|
|
|
| 250 |
)
|
| 251 |
|
| 252 |
|
| 253 |
+
def load_example_document(img, prompt, lang="ch"):
|
| 254 |
if img is not None:
|
| 255 |
document = prompt_files[prompt]
|
| 256 |
+
preview, answer, answer_text = process_prompt(prompt, document, lang)
|
| 257 |
return document, prompt, preview, gr.update(visible=True), answer, answer_text
|
| 258 |
else:
|
| 259 |
return None, None, None, gr.update(visible=False), None, None
|
|
|
|
| 355 |
with gr.Blocks(css=CSS) as demo:
|
| 356 |
gr.HTML(read_content("header.html"))
|
| 357 |
gr.Markdown(
|
| 358 |
+
f" ⚡DocPrompt⚡ is a Document Prompt Engine uses ERNIE-LayoutX as the backbone model.\n"
|
| 359 |
+
f" The engine is powered by Baidu Wenxin Document Intelligence Team 🚀 and is ability for multilingual documents information extraction and question ansering.\n"
|
|
|
|
| 360 |
)
|
| 361 |
|
| 362 |
document = gr.Variable()
|
|
|
|
| 462 |
|
| 463 |
prompt.submit(
|
| 464 |
fn=process_prompt,
|
| 465 |
+
inputs=[prompt, document, ocr_lang],
|
| 466 |
outputs=[image, output, output_text],
|
| 467 |
)
|
| 468 |
|
|
|
|
| 480 |
|
| 481 |
example_image.change(
|
| 482 |
fn=load_example_document,
|
| 483 |
+
inputs=[example_image, example_prompt, ocr_lang],
|
| 484 |
outputs=[document, prompt, image, img_clear_button, output, output_text],
|
| 485 |
)
|
| 486 |
|
|
|
|
| 487 |
gr.Markdown("[](https://github.com/PaddlePaddle/PaddleNLP)")
|
| 488 |
gr.HTML(read_content("footer.html"))
|
| 489 |
|
header.html
CHANGED
|
@@ -20,11 +20,7 @@
|
|
| 20 |
margin-bottom: 10px;
|
| 21 |
justify-content: center;
|
| 22 |
">
|
| 23 |
-
<a href="https://github.com/PaddlePaddle/PaddleNLP"><h1 style="font-weight: 900; margin-bottom: 7px;">
|
| 24 |
-
DocPrompt
|
| 25 |
</h1></a>
|
| 26 |
-
</div>
|
| 27 |
-
<!-- <p style="margin-bottom: 10px; font-weight: 900; font-size: 100%">
|
| 28 |
-
⚡DocPrompt⚡ is a Document Prompt Engine Powered by <br> Document Intelligence Technology in Baidu Wenxin<br>. 🚀 To use it, simply upload an image or PDF, type the prompt, and click 'submit', or click one of the examples to load them.
|
| 29 |
-
</p>
|
| 30 |
-
</div> -->
|
|
|
|
| 20 |
margin-bottom: 10px;
|
| 21 |
justify-content: center;
|
| 22 |
">
|
| 23 |
+
<a href="https://github.com/PaddlePaddle/PaddleNLP"><h1 style="font-weight: 900; align-items: center; margin-bottom: 7px;">
|
| 24 |
+
ERNIE-LayoutX DocPrompt Engine
|
| 25 |
</h1></a>
|
| 26 |
+
</div>
|
|
|
|
|
|
|
|
|
|
|
|
paddlenlp-preview.jpeg
DELETED
Git LFS Details
|