Spaces:
Running
Running
File size: 2,977 Bytes
29edf23 84669bc 6b18ba5 29edf23 30196dc 6b18ba5 936bfca 30196dc 936bfca c93f011 29edf23 c93f011 936bfca 6b18ba5 10dc1f6 30196dc b3aee5e 4d1390a ea28e08 6b18ba5 84669bc 6b18ba5 4d1390a 6b18ba5 776fa07 84669bc 4d1390a |
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 |
# Import dependencies
import gradio as gr
from transformers import AutoTokenizer, AutoModelForSequenceClassification, T5Tokenizer, T5ForConditionalGeneration
import torch
import nltk
from nltk.corpus import wordnet
import subprocess
# Download NLTK data (if not already downloaded)
nltk.download('punkt')
nltk.download('stopwords')
nltk.download('wordnet') # Download WordNet
# Check for GPU and set the device accordingly
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Load AI Detector model and tokenizer from Hugging Face (DistilBERT)
tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english").to(device)
# Load Parrot Paraphraser model and tokenizer for humanizing text
paraphrase_tokenizer = T5Tokenizer.from_pretrained("prithivida/parrot_paraphraser_on_T5")
paraphrase_model = T5ForConditionalGeneration.from_pretrained("prithivida/parrot_paraphraser_on_T5").to(device)
# AI detection function using DistilBERT
def detect_ai_generated(text):
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512).to(device)
with torch.no_grad():
outputs = model(**inputs)
probabilities = torch.softmax(outputs.logits, dim=1)
ai_probability = probabilities[0][1].item() # Probability of being AI-generated
return f"AI-Generated Content Probability: {ai_probability:.2f}%"
# Humanize the AI-detected text using the Parrot Paraphraser model
def humanize_text(AI_text):
inputs = paraphrase_tokenizer(AI_text, return_tensors="pt", max_length=512, truncation=True).to(device)
with torch.no_grad(): # Avoid gradient calculations for faster inference
paraphrased_ids = paraphrase_model.generate(
inputs['input_ids'],
max_length=inputs['input_ids'].shape[-1] + 20, # Slightly more than the original input length
num_beams=4,
early_stopping=True,
length_penalty=1.0,
no_repeat_ngram_size=3,
)
paraphrased_text = paraphrase_tokenizer.decode(paraphrased_ids[0], skip_special_tokens=True)
return f"Humanized Text:\n{paraphrased_text}"
# Gradio interface definition
ai_detection_interface = gr.Interface(
fn=detect_ai_generated,
inputs="textbox",
outputs="text",
title="AI Text Detection",
description="Enter text to determine the probability of it being AI-generated."
)
humanization_interface = gr.Interface(
fn=humanize_text,
inputs="textbox",
outputs="text",
title="Text Humanizer",
description="Enter text to get a human-written version, paraphrased for natural output."
)
# Combine both interfaces into a single Gradio app with tabs
interface = gr.TabbedInterface(
[ai_detection_interface, humanization_interface],
["AI Detection", "Humanization"]
)
# Launch the Gradio app
interface.launch(debug=False)
|