File size: 3,988 Bytes
ad96198 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 |
__all__ = ["app"]
import gradio as gr
import torch
import numpy as np
from transformers import AutoConfig, AutoTokenizer, DataCollatorWithPadding, DebertaV2ForSequenceClassification
MINIMUM_TOKENS = 48
FOUNDATION_MODEL_NAME = "binh230/deberta-base"
# Load the tokenizer and model for DeBERTa
tokenizer = AutoTokenizer.from_pretrained(FOUNDATION_MODEL_NAME)
config = AutoConfig.from_pretrained(FOUNDATION_MODEL_NAME)
config.num_labels = 2 # For binary classification
model = DebertaV2ForSequenceClassification.from_pretrained(FOUNDATION_MODEL_NAME, config=config)
model.to("cuda")
# Text processing and prediction function
def count_tokens(text):
return len(text.split())
def run_detector(input_str):
if count_tokens(input_str) < MINIMUM_TOKENS:
return f"Too short length. Need minimum {MINIMUM_TOKENS} tokens to run Binoculars."
# Tokenize input text
inputs = tokenizer(input_str, return_tensors="pt", padding=True, truncation=True).to("cuda")
# Run model and get prediction
with torch.no_grad():
outputs = model(**inputs)
logits = outputs.logits
prediction = torch.argmax(logits, dim=-1).item()
# Interpret prediction
return "Most likely AI-Generated" if prediction == 1 else "Most likely Human-Generated"
# Gradio app interface
css = """
.green { color: black!important; line-height:1.9em; padding: 0.2em 0.2em; background: #ccffcc; border-radius:0.5rem;}
.red { color: black!important; line-height:1.9em; padding: 0.2em 0.2em; background: #ffad99; border-radius:0.5rem;}
.hyperlinks {
display: flex;
align-items: center;
align-content: center;
padding-top: 12px;
justify-content: flex-end;
margin: 0 10px;
text-decoration: none;
color: #000;
}
"""
capybara_problem = '''Dr. Capy Cosmos, a capybara unlike any other, astounded the scientific community with his groundbreaking research...'''
with gr.Blocks(css=css, theme=gr.themes.Default(font=[gr.themes.GoogleFont("Inconsolata"), "Arial", "sans-serif"])) as app:
with gr.Row():
with gr.Column(scale=3):
gr.HTML("<h1>Mambaformer Detecting AI generated text</h1>")
with gr.Column(scale=1):
gr.HTML("""
<p>
<a href="https://github.com/DanielBinh2k3/Mamba-AI-generated-text-detection" target="_blank">code</a>
<a href="mailto:[email protected]" target="_blank">contact</a>
</p>
""", elem_classes="hyperlinks")
with gr.Row():
input_box = gr.Textbox(value=capybara_problem, placeholder="Enter text here", lines=8, label="Input Text")
with gr.Row():
submit_button = gr.Button("Run Detection", variant="primary")
clear_button = gr.ClearButton()
with gr.Row():
output_text = gr.Textbox(label="Prediction", value="Most likely AI-Generated")
with gr.Accordion("Disclaimer", open=False):
gr.Markdown("""
- `Accuracy`: AI-generated text detectors aim for accuracy, but no detector is perfect.
- `Use Cases`: This tool is most useful for detecting AI-generated content in moderation scenarios.
- `Known Weaknesses`: Non-English texts and highly memorized texts (like constitutions) may yield unreliable results.
""")
with gr.Accordion("Cite our work", open=False):
gr.Markdown("""
```bibtex
@misc{BamBa2024llm,
title={Enhancing AI Text Detection through MambaFormer and Adversarial Learning Techniques},
author={Truong Nguyen Gia Binh},
year={2024},
eprint={},
archivePrefix={},
primaryClass={}
}
```
""")
submit_button.click(run_detector, inputs=input_box, outputs=output_text)
clear_button.click(lambda: ("", ""), outputs=[input_box, output_text])
# Run the Gradio app
if __name__ == "__main__":
app.launch(share=True)
|