【大模型】Unsloth安装及使用教程

news2024/9/22 5:42:26

Unsloth是一个开源的大模型训练加速项目，使用OpenAI的Triton对模型的计算过程进行重写，大幅提升模型的训练速度，降低训练中的显存占用。Unsloth能够保证重写后的模型计算的一致性，实现中不存在近似计算，模型训练的精度损失为零。Unsloth支持绝大多数主流的GPU设备，包括V100, T4, Titan V, RTX 20, 30, 40x, A100, H100, L40等，支持对LoRA和QLoRA的训练加速和高效显存管理，支持Flash Attention。

Unsloth Github项目：https://github.com/unslothai/unsloth

在线使用

Unsloth支持主流的Llama、Mistral、Gemma等大模型，且对上述模型的微调性能级内存占用有显著提升，如下图所示：

通过点击“Start for free”按钮，可以启动Google的colab笔记本，这里Google为新账号免费提供几个小时的云端GPU服务器资源（Tesla T4）资源供大家使用。在colab笔记本训练得到的微调模型可以导出为GGUF格式，并上传至Hugging Face网站上。

以Llama 3.1(8B)为例，Google的colab网址为：https://colab.research.google.com/drive/1Ys44kVvmeZtnICzWz0xgpRnrIOjZAuxp?usp=sharing

colab中给出了在线运行的代码，我们配置好基础环境后只需按序执行代码即可。

from unsloth import FastLanguageModel
import torch
max_seq_length = 2048 # Choose any! We auto support RoPE Scaling internally!
dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.

# 4bit pre quantized models we support for 4x faster downloading + no OOMs.
fourbit_models = [
    "unsloth/Meta-Llama-3.1-8B-bnb-4bit",      # Llama-3.1 15 trillion tokens model 2x faster!
    "unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit",
    "unsloth/Meta-Llama-3.1-70B-bnb-4bit",
    "unsloth/Meta-Llama-3.1-405B-bnb-4bit",    # We also uploaded 4bit for 405b!
    "unsloth/Mistral-Nemo-Base-2407-bnb-4bit", # New Mistral 12b 2x faster!
    "unsloth/Mistral-Nemo-Instruct-2407-bnb-4bit",
    "unsloth/mistral-7b-v0.3-bnb-4bit",        # Mistral v3 2x faster!
    "unsloth/mistral-7b-instruct-v0.3-bnb-4bit",
    "unsloth/Phi-3-mini-4k-instruct",          # Phi-3 2x faster!d
    "unsloth/Phi-3-medium-4k-instruct",
    "unsloth/gemma-2-9b-bnb-4bit",
    "unsloth/gemma-2-27b-bnb-4bit",            # Gemma 2x faster!
] # More models at https://huggingface.co/unsloth

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/Meta-Llama-3.1-8B",
    max_seq_length = max_seq_length,
    dtype = dtype,
    load_in_4bit = load_in_4bit,
    # token = "hf_...", # use one if using gated models like meta-llama/Llama-2-7b-hf
)