app.py

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

# Model initialization
model_name = "gpt2"  # or your specific model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Add special tokens for languages
special_tokens = {"additional_special_tokens": ["[YORUBA]", "[IGBO]", "[HAUSA]"]}
tokenizer.add_special_tokens(special_tokens)
model.resize_token_embeddings(len(tokenizer))

# Move model to GPU if available
device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)

def generate_text(prompt, language):
    # Add language tag to prompt
    tagged_prompt = f"[{language.upper()}] {prompt}"
    
    # Tokenize
    inputs = tokenizer(tagged_prompt, return_tensors="pt").to(device)
    
    # Generate
    outputs = model.generate(
        **inputs,
        max_length=100,
        num_return_sequences=1,
        temperature=0.7,
        pad_token_id=tokenizer.eos_token_id
    )
    
    # Decode and return
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# Create Gradio interface
iface = gr.Interface(
    fn=generate_text,
    inputs=[
        gr.Textbox(label="Enter your prompt"),
        gr.Dropdown(choices=["YORUBA", "IGBO", "HAUSA"], label="Select Language")
    ],
    outputs=gr.Textbox(label="Generated Text"),
    title="Nigerian Language Generator",
    description="Generate text in Yoruba, Igbo, or Hausa using a fine-tuned GPT model."
)

if __name__ == "__main__":
    iface.launch()