app.py

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer, pipeline
from threading import Thread
import gradio as gr

DEVICE = "cpu"
if torch.cuda.is_available():
  DEVICE = "cuda"

# The huggingface model id for phi-2 instruct model
checkpoint = "rasyosef/phi-2-instruct-v0.1"

# Download and load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(
    checkpoint,
    torch_dtype=torch.float32,
    device_map=DEVICE
  )


# Text generation pipeline
phi2 = pipeline(
    "text-generation",
    tokenizer=tokenizer,
    model=model,
    pad_token_id=tokenizer.eos_token_id,
    eos_token_id=[tokenizer.eos_token_id],
    device_map=DEVICE
)

# Function that accepts a prompt and generates text using the phi2 pipeline
def generate(message, chat_history, max_new_tokens=64):

  history = [
      {"role": "system", "content": "You are Phi, a helpful AI assistant made by Microsoft and RasYosef. User will you give you a task. Your goal is to complete the task as faithfully as you can."}
  ]

  for sent, received in chat_history:
    history.append({"role": "user", "content": sent})
    history.append({"role": "assistant", "content": received})

  history.append({"role": "user", "content": message})
  #print(history)

  if len(tokenizer.apply_chat_template(history)) > 512:
    yield "chat history is too long"
  else:
    # Streamer
    streamer = TextIteratorStreamer(tokenizer=tokenizer, skip_prompt=True, skip_special_tokens=True, timeout=300.0)
    thread = Thread(target=phi2, kwargs={"text_inputs":history, "max_new_tokens":max_new_tokens, "streamer":streamer})
    thread.start()

    generated_text = ""
    for word in streamer:
      generated_text += word
      response = generated_text.strip()

      yield response

# Chat interface with gradio
with gr.Blocks() as demo:
  gr.Markdown("""
  # Phi-2 Chat Demo
  This chatbot was created using a finetuned version of Microsoft's 2.7 billion parameter Phi 2 transformer model, [phi-2-instruct-v0.1](https://huggingface.co/rasyosef/phi-2-instruct-v0.1) that has underwent a post-training process that incorporates both **supervised fine-tuning** and **direct preference optimization** for instruction following.
  """)

  tokens_slider = gr.Slider(8, 256, value=64, label="Maximum new tokens", info="A larger `max_new_tokens` parameter value gives you longer text responses but at the cost of a slower response time.")

  chatbot = gr.ChatInterface(
    chatbot=gr.Chatbot(height=400),
    fn=generate,
    additional_inputs=[tokens_slider],
    stop_btn=None,
    examples=[
        ["Hi"],
        ["What's the German word for 'car'?"],
        ["Molly and Abigail want to attend a beauty and modeling contest. They both want to buy new pairs of shoes and dresses. Molly buys a pair of shoes which costs $40 and a dress which costs $160. How much should Abigail budget if she wants to spend half of what Molly spent on the pair of shoes and dress?"],
      ]
  )

demo.queue().launch(debug=True)