news-class-1 / app.py
zionia's picture
edit format
d0c2053 verified
import gradio as gr
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
import csv
MODEL_URL = "https://huggingface.co/dsfsi/PuoBERTa-News"
WEBSITE_URL = "https://www.kodiks.com/ai_solutions.html"
tokenizer = AutoTokenizer.from_pretrained("dsfsi/PuoBERTa-News")
model = AutoModelForSequenceClassification.from_pretrained("dsfsi/PuoBERTa-News")
categories = {
"arts_culture_entertainment_and_media": "Botsweretshi, setso, boitapoloso le bobegakgang",
"crime_law_and_justice": "Bosenyi, molao le bosiamisi",
"disaster_accident_and_emergency_incident": "Masetlapelo, kotsi le tiragalo ya maemo a tshoganyetso",
"economy_business_and_finance": "Ikonomi, tsa kgwebo le tsa ditšhelete",
"education": "Thuto",
"environment": "Tikologo",
"health": "Boitekanelo",
"politics": "Dipolotiki",
"religion_and_belief": "Bodumedi le tumelo",
"society": "Setšhaba"
}
def prediction(news):
classifier = pipeline("text-classification", tokenizer=tokenizer, model=model, return_all_scores=True)
preds = classifier(news)
preds_dict = {categories.get(pred['label'], pred['label']): round(pred['score'], 4) for pred in preds[0]}
return preds_dict
def file_prediction(file):
news_list = []
if file.name.endswith('.csv'):
file.seek(0)
reader = csv.reader(file.read().decode('utf-8').splitlines())
news_list = [row[0] for row in reader if row]
else:
file.seek(0)
file_content = file.read().decode('utf-8')
news_list = file_content.splitlines()
results = []
for news in news_list:
if news.strip():
pred = prediction(news)
results.append([news, pred])
return results
with gr.Blocks() as demo:
with gr.Row():
with gr.Column(scale=1):
pass
with gr.Column(scale=4, min_width=1000):
gr.Image("logo_transparent_small.png", elem_id="logo", show_label=False, width=500)
gr.Markdown("""
<h1 style='text-align: center;'>Setswana News Classification</h1>
<p style='text-align: center;'>This space provides a classification service for news in Setswana.</p>
""")
with gr.Column(scale=1):
pass
with gr.Tabs():
with gr.Tab("Text Input"):
gr.Markdown(f"""
Enter Setswana news article to see the category of the news. <br>
For this classification, the <a href='{MODEL_URL}' target='_blank'>PuoBERTa-News</a> model was used.
""")
inp_text = gr.Textbox(lines=10, label="Paste some Setswana news here")
output_label = gr.Label(num_top_classes=5, label="News categories probabilities")
translate_button = gr.Button("Classify")
translate_button.click(prediction, inputs=inp_text, outputs=output_label)
with gr.Tab("File Upload"):
gr.Markdown("""
Upload a text or CSV file with Setswana news articles. The first column in the CSV should contain the news text.
""")
file_input = gr.File(label="Upload text or CSV file")
file_output = gr.Dataframe(headers=["News Text", "Category Predictions"], label="Predictions from file")
file_button = gr.Button("Classify File")
file_button.click(file_prediction, inputs=file_input, outputs=file_output)
gr.Markdown("""
<div style='text-align: center;'>
<a href='https://github.com/dsfsi/PuoBERTa-News' target='_blank'>GitHub</a> |
<a href='https://docs.google.com/forms/d/e/1FAIpQLSf7S36dyAUPx2egmXbFpnTBuzoRulhL5Elu-N1eoMhaO7v10w/viewform' target='_blank'>Feedback Form</a>
</div>
""")
with gr.Accordion("More Information", open=False):
gr.Markdown("""
<h4 style="text-align: center;">Authors</h4>
<div style='text-align: center;'>
Vukosi Marivate, Moseli Mots'Oehli, Valencia Wagner, Richard Lastrucci, Isheanesu Dzingirai
</div>
""")
gr.Markdown("""
<h4 style="text-align: center;">Citation</h4>
<pre style="text-align: left; white-space: pre-wrap;">
@inproceedings{marivate2023puoberta,
title = {PuoBERTa: Training and evaluation of a curated language model for Setswana},
author = {Vukosi Marivate and Moseli Mots'Oehli and Valencia Wagner and Richard Lastrucci and Isheanesu Dzingirai},
year = {2023},
booktitle= {Artificial Intelligence Research. SACAIR 2023. Communications in Computer and Information Science},
url= {https://link.springer.com/chapter/10.1007/978-3-031-49002-6_17},
keywords = {NLP},
preprint_url = {https://arxiv.org/abs/2310.09141},
dataset_url = {https://github.com/dsfsi/PuoBERTa},
software_url = {https://huggingface.co/dsfsi/PuoBERTa}
}
</pre>
""")
gr.Markdown("""
<h4 style="text-align: center;">DOI</h4>
<div style='text-align: center;'>
DOI: <a href="https://doi.org/10.1007/978-3-031-49002-6_17" target="_blank">10.1007/978-3-031-49002-6_17</a>
</div>
""")
demo.launch()