app.py

import gradio as gr
from transformers import RobertaTokenizer, RobertaForSequenceClassification
import torch

# Define available models
model_options = {
    "GoalZero/aidetection-ada-v0.2": "GoalZero/aidetection-ada-v0.2",
    "GoalZero/aidetection-ada-v0.1": "GoalZero/aidetection-ada-v0.1",
    "GoalZero/babbage-mini-v0.1": "GoalZero/babbage-mini-v0.1"
}

# Initialize global variables for model and tokenizer
model = None
tokenizer = None

def load_model(model_name):
    """Helper function to load model and tokenizer"""
    try:
        return (
            RobertaForSequenceClassification.from_pretrained(model_name),
            RobertaTokenizer.from_pretrained(model_name)
        )
    except Exception as e:
        raise Exception(f"Failed to load model {model_name}: {str(e)}")

# Load default model
try:
    default_model = "GoalZero/aidetection-ada-v0.2"
    model, tokenizer = load_model(default_model)
except Exception as e:
    print(f"Error loading default model: {str(e)}")

def classify_text(text, model_choice):
    global model, tokenizer
    
    try:
        # Check if we need to change the model
        if model is None or model_choice != model.name_or_path:
            model, tokenizer = load_model(model_choice)
        
        # Clean the input text
        cleaned_text = text.replace('.', '').replace('\n', ' ')
        
        # Tokenize the cleaned input text
        inputs = tokenizer(
            cleaned_text,
            return_tensors='pt',
            padding=True,
            truncation=True,
            max_length=128
        )
        
        # Get the model's prediction
        with torch.no_grad():
            outputs = model(**inputs)
            
            # Apply softmax to get probabilities
            probabilities = torch.nn.functional.softmax(outputs.logits, dim=-1)
            
            # Get the probability of class '1'
            prob_1 = probabilities[0][1].item()
            
            return {
                "AI Probability": round(prob_1 * 100, 10),
                "Model used": model_choice
            }
            
    except Exception as e:
        return {
            "error": f"An error occurred: {str(e)}",
            "Model used": model_choice
        }

# Create the Gradio interface
iface = gr.Interface(
    fn=classify_text,
    inputs=[
        gr.Textbox(
            lines=2,
            placeholder="Enter text here...",
            label="Input Text"
        ),
        gr.Dropdown(
            choices=list(model_options.keys()),
            value="GoalZero/aidetection-ada-v0.2",
            label="Select Model Version"
        )
    ],
    outputs=gr.JSON(label="Results"),
    title="GoalZero Ada AI Detection",
    description="Enter text to get the probability of it being AI-written. Select a model version to use.",
    examples=[
        ["Waymo is an American autonomous driving technology company that originated as the Google Self-Driving Car Project in 2009. It is now a subsidiary of Alphabet Inc., headquartered in Mountain View, California. The name \"Waymo\" was adopted in December 2016 when the project was rebranded and spun out of Google to focus on developing fully autonomous vehicles aimed at improving transportation safety and convenience", "GoalZero/babbage-mini-v0.1"],
        ["WWII demonstrated the importance of alliances in global conflicts. The Axis and Allied powers were formed as countries sought to protect their interests and expand their influence. This lesson underscores the potential for future global conflicts to involve complex alliances, similar to the Cold War era’s NATO and Warsaw Pact alignments.", "GoalZero/aidetection-ada-v0.2"],
        ["Eustace was a thorough gentleman. There was candor in his quack, and affability in his waddle; and underneath his snowy down beat a pure and sympathetic heart. In short, he was a most exemplary duck.", "GoalZero/aidetection-ada-v0.1"]
    ]
)

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