app.py

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

# Load the model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("unsloth/Llama-3.2-1B")
model = AutoModelForCausalLM.from_pretrained("unsloth/Llama-3.2-1B")

# Use a pipeline for text generation
text_gen_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)

# Text generation function with additional parameters
def generate_text(prompt, max_length=50, temperature=0.7, top_p=0.9, top_k=50):
    generated_text = text_gen_pipeline(prompt, 
                                       max_length=max_length, 
                                       temperature=temperature, 
                                       top_p=top_p, 
                                       top_k=top_k, 
                                       num_return_sequences=1)
    return generated_text[0]['generated_text']

# Gradio Interface
with gr.Blocks() as demo:
    gr.Markdown("## Text Generation with Llama 3.2 - 1B")
    
    # Input box for user prompt
    prompt_input = gr.Textbox(label="Input (Prompt)", placeholder="Enter your prompt here...")
    
    # Slider for maximum text length
    max_length_input = gr.Slider(minimum=10, maximum=200, value=50, step=10, label="Maximum Length")
    
    # Slider for temperature (controls creativity)
    temperature_input = gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature (creativity)")
    
    # Slider for top_p (nucleus sampling)
    top_p_input = gr.Slider(minimum=0.1, maximum=1.0, value=0.9, step=0.1, label="Top-p (nucleus sampling)")
    
    # Slider for top_k (controls diversity)
    top_k_input = gr.Slider(minimum=1, maximum=100, value=50, step=1, label="Top-k (sampling diversity)")
    
    # Output box for the generated text
    output_text = gr.Textbox(label="Generated Text")
    
    # Submit button
    generate_button = gr.Button("Generate")
    
    # Action on button click
    generate_button.click(generate_text, inputs=[prompt_input, max_length_input, temperature_input, top_p_input, top_k_input], outputs=output_text)

# Launch the app
demo.launch()