app.py

import torch
from transformers import DistilBertTokenizer, DistilBertForSequenceClassification
import gradio as gr
import re
import nltk
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer

# Download necessary NLTK data
nltk.download('punkt')
nltk.download('stopwords')
nltk.download('wordnet')

# Preprocessing function
def preprocess(text):
    text = re.sub(r'[^a-zA-Z\s]', '', text).lower()
    tokens = word_tokenize(text)
    stop_words = set(stopwords.words('english'))
    tokens = [word for word in tokens if word not in stop_words]
    lemmatizer = WordNetLemmatizer()
    tokens = [lemmatizer.lemmatize(word) for word in tokens]
    return ' '.join(tokens)

# Load tokenizer and model
tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased')
model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased', num_labels=2)
model.load_state_dict(torch.load('best_model (3).pth', map_location=torch.device('cpu')))
model.eval()

# Prediction function
def classify_essay(text):
    cleaned_text = preprocess(text)
    inputs = tokenizer(cleaned_text, return_tensors='pt', truncation=True, padding=True, max_length=100)
    with torch.no_grad():
        outputs = model(**inputs)
        probs = torch.nn.functional.softmax(outputs.logits, dim=1)
        predicted_class = torch.argmax(probs, dim=1).item()
        labels = ["Human-Written", "AI-Generated"]
        return {labels[0]: float(probs[0][0]), labels[1]: float(probs[0][1])}

# Gradio interface
iface = gr.Interface(
    fn=classify_essay,
    inputs=gr.Textbox(lines=10, placeholder="Paste your essay here..."),
    outputs=gr.Label(num_top_classes=2),
    title="Essay Authorship Classifier",
    description="Detect whether an essay is AI-generated or human-written using a fine-tuned DistilBERT model."
)

# Launch the app
if __name__ == "__main__":
    iface.launch()