如何使用 Code Llama 🤖 构建自己的 LLM 编码助手（含代码）

import torchfrom transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
class ChatModel:def __init__(self, model="codellama/CodeLlama-7b-Instruct-hf"):quantization_config = BitsAndBytesConfig(load_in_4bit=True, # use 4-bit quantizationbnb_4bit_compute_dtype=torch.float16,bnb_4bit_use_double_quant=True,)self.model = AutoModelForCausalLM.from_pretrained(model,quantization_config=quantization_config,device_map="cuda",cache_dir="./models", # download model to the models folder)self.tokenizer = AutoTokenizer.from_pretrained(model, use_fast=True, padding_side="left")

self.history = []self.history_length = 1

self.DEFAULT_SYSTEM_PROMPT = """\\You are a helpful, respectful and honest assistant with a deep knowledge of code and software design. Always answer as helpfully as possible, while being safe. Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.\\n\\nIf a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.\\"""

def append_to_history(self, user_prompt, response):self.history.append((user_prompt, response))if len(self.history) > self.history_length:self.history.pop(0)

def generate(self, user_prompt, system_prompt, top_p=0.9, temperature=0.1, max_new_tokens=512):
texts = [f"<s>[INST] <<SYS>>\\n{system_prompt}\\n<</SYS>>\\n\\n"]do_strip = Falsefor old_prompt, old_response in self.history:old_prompt = old_prompt.strip() if do_strip else old_promptdo_strip = Truetexts.append(f"{old_prompt} [/INST] {old_response.strip()} </s><s>[INST] ")user_prompt = user_prompt.strip() if do_strip else user_prompttexts.append(f"{user_prompt} [/INST]")prompt = "".join(texts)
inputs = self.tokenizer(prompt, return_tensors="pt", add_special_tokens=False).to("cuda")
output = self.model.generate(inputs["input_ids"],attention_mask=inputs["attention_mask"],pad_token_id=self.tokenizer.eos_token_id,max_new_tokens=max_new_tokens,do_sample=True,top_p=top_p,top_k=50,temperature=temperature,)output = output[0].to("cpu")response = self.tokenizer.decode(output[inputs["input_ids"].shape[1] : -1])self.append_to_history(user_prompt, response)return response

from ChatModel import *
model = ChatModel()response = model.generate(user_prompt="Write a hello world program in C++",system_prompt=model.DEFAULT_SYSTEM_PROMPT)print(response)
 Sure, here is a simple "Hello World" program in C++:
#include <iostream>int main() {std::cout << "Hello, World!" << std::endl;return 0;}This program will print "Hello, World!" to the console when it is run.The `std::cout` statement is used to print the message to the console, and the `std::endl` statement is used to print a newline character after the message.The `return 0;` statement is used to indicate that the program has completed successfully.

import streamlit as stfrom ChatModel import *
st.title("Code Llama Assistant")
@st.cache_resourcedef load_model():model = ChatModel()return model
model = load_model()# load our ChatModel once and then cache it

with st.sidebar:temperature = st.slider("temperature", 0.0, 2.0, 0.1)top_p = st.slider("top_p", 0.0, 1.0, 0.9)max_new_tokens = st.number_input("max_new_tokens", 128, 4096, 256)system_prompt = st.text_area("system prompt", value=model.DEFAULT_SYSTEM_PROMPT, height=500)

# Initialize chat historyif "messages" not in st.session_state:st.session_state.messages = []
# Display chat messages from history on app rerunfor message in st.session_state.messages:with st.chat_message(message["role"]):st.markdown(message["content"])
# Accept user inputif prompt := st.chat_input("Ask me anything!"):# Add user message to chat historyst.session_state.messages.append({"role": "user", "content": prompt})# Display user message in chat message containerwith st.chat_message("user"):st.markdown(prompt)
# Display assistant response in chat message containerwith st.chat_message("assistant"):user_prompt = st.session_state.messages[-1]["content"]answer = model.generate(user_prompt,top_p=top_p,temperature=temperature,max_new_tokens=max_new_tokens,system_prompt=system_prompt,)response = st.write(answer)st.session_state.messages.append({"role": "assistant", "content": answer})