app.py

import torch
from transformers import AutoProcessor, AutoModelForImageTextToText
from peft import PeftModel
import gradio as gr

# Set up device (CPU or GPU)
device = "cuda" if torch.cuda.is_available() else "cpu"

# Load processor and model
model_name = "adarsh3601/my_gemma3_pt"  # Change to your model path
processor = AutoProcessor.from_pretrained(model_name)
model = AutoModelForImageTextToText.from_pretrained(model_name).to(device)

# Optional: If using PEFT model with adapter
# adapter_model_id = "your_adapter_model_id"  # Uncomment and replace if using adapter
# model = PeftModel.from_pretrained(model, adapter_model_id)

# Define function to process the user input
def chat(prompt):
    # Prepare the message in the format the model expects
    messages = [{"role": "user", "content": prompt}]
    
    # Process the input using the processor
    inputs = processor(messages, return_tensors="pt").to(device)
    
    # Generate the output from the model
    with torch.no_grad():
        outputs = model.generate(**inputs, max_length=200)

    # Decode and return the response
    return processor.decode(outputs[0], skip_special_tokens=True)

# Gradio interface
gr.Interface(
    fn=chat,
    inputs="text",
    outputs="text",
    title="Gemma Chat Model",
    description="Chat with Gemma3 model",
    live=True
).launch(share=False)  # share=False for Hugging Face Spaces