app.py

import os
import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

# ------------------------------
# Load model
# ------------------------------
#model_id = "gemma_3_270m_model"   # your model folder or HF repo
model_id = "google/gemma-3-270m"   # your model folder or HF repo
hf_token = os.environ.get("HF_TOKEN")  # read from Hugging Face Secrets

tokenizer = AutoTokenizer.from_pretrained(
    model_id,
    use_auth_token=hf_token,
    trust_remote_code=True
)

model = AutoModelForCausalLM.from_pretrained(
    model_id,
    use_auth_token=hf_token,
    trust_remote_code=True,
    device_map="auto"
)

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer
)

# ------------------------------
# Gradio interface
# ------------------------------
def generate_text(prompt, max_length=100):
    """Generate text from the model"""
    output = pipe(prompt, max_length=max_length)
    return output[0]['generated_text']

# Create Gradio interface
demo = gr.Interface(
    fn=generate_text,
    inputs=[
        gr.Textbox(label="Prompt", placeholder="Enter your text here..."),
        gr.Slider(label="Max length", minimum=10, maximum=500, value=100)
    ],
    outputs=gr.Textbox(label="Generated Text"),
    title="Gemma-3-270M Text Generator",
    description="Enter a prompt and the model will generate text."
)

# Launch the app
if __name__ == "__main__":
    demo.launch()