app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = 'deepseek-ai/deepseek-coder-33b-instruct'

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

def call_llm(input_text):
    input_ids = tokenizer.encode(input_text, return_tensors="pt")
    
    kwargs = {
        "max_length": 500,
        "num_return_sequences": 1,
        "temperature": 0.7,
        "top_k": 50
    }
    
    # Generate text
    output_ids = model.generate(input_ids, **kwargs)
    
    # Decode and print the output
    output_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
    print(output_text)
    return output_text

with gr.Blocks() as app:
    chat = gr.ChatInterface(
        call_llm,
            )
app.launch()