nothashim/CodeSage

1import gradio as gr2from transformers import T5ForConditionalGeneration, T5Tokenizer3import torch4 5model_name = "nothashim/CodeSage"6model = T5ForConditionalGeneration.from_pretrained(model_name)7tokenizer = T5Tokenizer.from_pretrained(model_name)8 9def generate_text(input_text, max_length=50, temperature=0.7):10    inputs = tokenizer(input_text, return_tensors="pt")11    with torch.no_grad():12        outputs = model.generate(13            **inputs,14            max_length=max_length,15            temperature=temperature,16            do_sample=True,17            pad_token_id=tokenizer.pad_token_id,18            eos_token_id=tokenizer.eos_token_id19        )20    return tokenizer.decode(outputs[0], skip_special_tokens=True)21 22iface = gr.Interface(23    fn=generate_text,24    inputs=[25        gr.Textbox(label="Input Text"),26        gr.Slider(minimum=10, maximum=100, value=50, step=1, label="Max Length"),27        gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.1, label="Temperature")28    ],29    outputs=gr.Textbox(label="Generated Text"),30    title="CodeSaga Transformer",31    description="A custom encoder-decoder transformer model for text generation."32)33 34if __name__ == "__main__":35    iface.launch()36