app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Load the pre-trained model and tokenizer
model_name = "khanfs/ChemSolubilityBERTa"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# Define the prediction function
def predict_solubility(smiles_string):
    inputs = tokenizer(smiles_string, return_tensors='pt', truncation=True, padding='max_length', max_length=128)
    with torch.no_grad():
        outputs = model(**inputs)
    solubility = outputs.logits.item()
    return f"Predicted Solubility: {solubility:.4f} log mol/L"

# Gradio interface
iface = gr.Interface(
    fn=predict_solubility,
    inputs="text",
    outputs="text",
    title="ChemSolubilityBERTa",
    description="Enter a SMILES string to predict its aqueous solubility using ChemSolubilityBERTa.",
    examples=[["CCO"], ["CC(C)=O"], ["C1=CC=CC=C1"]]  # Example SMILES strings for ethanol, acetone, and benzene
)

# Launch the app
iface.launch()