app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model and tokenizer
model_name = "deepseek-ai/DeepSeek-V3"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)

# Function to handle chatbot response
def chat_with_deepseek(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(
        inputs["input_ids"],
        max_length=512,
        num_return_sequences=1,
        pad_token_id=tokenizer.eos_token_id
    )
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response

# Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# DeepSeek Chatbot")
    with gr.Row():
        with gr.Column():
            user_input = gr.Textbox(label="Enter your message", placeholder="Type something...")
        with gr.Column():
            submit_btn = gr.Button("Send")
    chatbot_output = gr.Textbox(label="Response", placeholder="Chatbot response will appear here")
    
    submit_btn.click(chat_with_deepseek, inputs=user_input, outputs=chatbot_output)

demo.launch()