app.py

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

model_name = "deepseek-ai/deepseek-coder-1.3b-base"

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)
model.eval()

def respond(prompt, max_tokens, temperature, top_p):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(
        **inputs,
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id,
    )
    return tokenizer.decode(outputs[0], skip_special_tokens=True)[len(prompt):].strip()

with gr.Blocks() as demo:
    gr.Markdown("# DeepSeek Coder without Login")
    prompt = gr.Textbox(label="Enter your prompt", lines=5)
    max_tokens = gr.Slider(1, 1024, value=512, step=1, label="Max Tokens")
    temperature = gr.Slider(0.1, 1.0, value=0.7, step=0.05, label="Temperature")
    top_p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p")
    btn = gr.Button("Generate")
    output = gr.Textbox(label="Output", lines=15)
    btn.click(respond, inputs=[prompt, max_tokens, temperature, top_p], outputs=output)

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