app.py

app_code = """\
import gradio as gr
import torch
import torch.nn.functional as F
from transformers import AutoModelForSequenceClassification, AutoTokenizer

# Load the fine-tuned SBERT model from Hugging Face
model_name = "Steph974/SBERT-FineTuned-Classifier"  # Your uploaded model
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

# Ensure the model is on the correct device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
model.eval()

def predict_similarity(sentence1, sentence2):
    \"\"\"
    Predicts the probability of two sentences belonging to the same class (1) or different (0).
    Returns probability instead of class label.
    \"\"\"
    # Tokenize input
    inputs = tokenizer(sentence1, sentence2, return_tensors="pt", truncation=True, padding="max_length", max_length=512)
    inputs = {key: value.to(device) for key, value in inputs.items()}  # Move tensors to model device

    # Perform inference
    with torch.no_grad():
        outputs = model(**inputs)

    # Get probabilities
    probabilities = F.softmax(outputs.logits, dim=1).cpu().numpy()[0]
    proba_same = probabilities[1]  # Probability that sentences are in the same class
    proba_diff = probabilities[0]  # Probability that sentences are different

    return {
        "Same Class Probability": round(proba_same * 100, 2),
        "Different Class Probability": round(proba_diff * 100, 2)
    }

# Gradio UI
interface = gr.Interface(
    fn=predict_similarity,
    inputs=[
        gr.Textbox(label="Sentence 1", placeholder="Enter the first sentence..."),
        gr.Textbox(label="Sentence 2", placeholder="Enter the second sentence...")
    ],
    outputs=gr.Label(label="Prediction Probabilities"),
    title="SBERT Sentence-Pair Similarity",
    description="Enter two sentences and see how similar they are according to the fine-tuned SBERT model.",
    theme="huggingface",
)

# Launch the Gradio app
interface.launch()
"""

# Save to app.py
with open("app.py", "w") as f:
    f.write(app_code)