app.py

from transformers import AutoTokenizer, AutoModelForCausalLM
import os
from huggingface_hub import login
import gradio as gr

# 1. Authenticate with Hugging Face token from secrets
hf_token = os.environ.get("hf_space_token")
login(token=hf_token)

# 2. Load Gemma model and tokenizer (GATED model needs token)
model_name = "google/gemma-3-1b-it"
tokenizer = AutoTokenizer.from_pretrained(model_name, token=hf_token)
model = AutoModelForCausalLM.from_pretrained(model_name, token=hf_token)

# 3. Define response generation function
def generate_response(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(**inputs, max_new_tokens=100, do_sample=True, temperature=0.7)
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# 4. Create Gradio interface
iface = gr.Interface(
    fn=generate_response,
    inputs=gr.Textbox(lines=2, placeholder="Ask something..."),
    outputs="text",
    title="Chat with Gemma",
    description="This chatbot is powered by Google's Gemma model running in Hugging Face Spaces."
)

# 5. Launch app
iface.launch()