371 lines
14 KiB
Markdown
371 lines
14 KiB
Markdown
|
# LLaMA Efficient Tuning
|
|||
|
|
|||
|
[![GitHub Repo stars](https://img.shields.io/github/stars/hiyouga/LLaMA-Efficient-Tuning?style=social)](https://github.com/hiyouga/LLaMA-Efficient-Tuning/stargazers)
|
|||
|
[![GitHub Code License](https://img.shields.io/github/license/hiyouga/LLaMA-Efficient-Tuning)](LICENSE)
|
|||
|
[![GitHub last commit](https://img.shields.io/github/last-commit/hiyouga/LLaMA-Efficient-Tuning)](https://github.com/hiyouga/LLaMA-Efficient-Tuning/commits/main)
|
|||
|
[![PyPI](https://img.shields.io/pypi/v/llmtuner)](https://pypi.org/project/llmtuner/)
|
|||
|
[![GitHub pull request](https://img.shields.io/badge/PRs-welcome-blue)](https://github.com/hiyouga/LLaMA-Efficient-Tuning/pulls)
|
|||
|
|
|||
|
👋 加入我们的 [微信群](assets/wechat.jpg).
|
|||
|
|
|||
|
\[ [English](README.md) | 中文 \]
|
|||
|
|
|||
|
## 更新日志
|
|||
|
|
|||
|
[23/07/19] 现在我们在该仓库中提供了对于 **LLaMA-2** 模型的训练支持. 试试 `--model_name_or_path meta-llama/Llama-2-7b-hf` 参数来使用 LLaMA-2 模型. 使用 LLaMA-2-chat 模型时记得使用 `--prompt_template llama2` 参数.
|
|||
|
|
|||
|
[23/07/18] 我们开发了一个用于训练、评估和推理的 all-in-one Web UI, . 试试 `train_web.py` 在浏览器中微调模型. 感谢 [@KanadeSiina](https://github.com/KanadeSiina) 和 [@codemayq](https://github.com/codemayq) 在项目发展中做出的努力.
|
|||
|
|
|||
|
[23/07/11] 现在我们在该仓库中提供了对于 **Baichuan-13B** 模型的训练支持. 训练 Baichuan-13B 时使用 `tests/modeling_baichuan.py` , 然后试试 `--model_name_or_path path_to_baichuan_model` 和 `--lora_target W_pack` 参数来训练 Baichuan-13B 模型. 使用 Baichuan-13B-Chat 模型时记得使用 `--prompt_template baichuan` 参数.
|
|||
|
|
|||
|
[23/07/09] 我们开源了 [FastEdit](https://github.com/hiyouga/FastEdit)⚡🩹,一个简单易用的、能迅速编辑大模型事实记忆的工具包。如果您感兴趣请关注我们的 [FastEdit](https://github.com/hiyouga/FastEdit) 项目。
|
|||
|
|
|||
|
[23/07/07] 现在我们在该仓库中提供了对于 **InternLM-7B** 模型的训练支持. 试试 `--model_name_or_path internlm/internlm-7b` 参数来使用 InternLM 模型. 使用 InternLM-chat 模型时记得使用 `--prompt_template intern` 参数.
|
|||
|
|
|||
|
[23/07/05] 现在我们在该仓库中提供了对于 **Falcon-7B/40B** 模型的训练支持. 试试 `--model_name_or_path tiiuae/falcon-7b` 和 `--lora_target query_key_value` 参数来使用 Falcon 模型.
|
|||
|
|
|||
|
[23/06/29] 我们提供了一个使用 instruction-following 数据集训练聊天模型的 **可复现的示例** , 更多细节请看这里 [Hugging Face Repo](https://huggingface.co/hiyouga/baichuan-7b-sft).
|
|||
|
|
|||
|
[23/06/22] 我们对齐了[示例 API](src/api_demo.py) 与 [OpenAI API](https://platform.openai.com/docs/api-reference/chat) 的格式,您可以将微调模型接入任意基于 ChatGPT 的应用中.
|
|||
|
|
|||
|
[23/06/15] 现在我们在该仓库中提供了对于 **Baichuan-7B** 模型的训练支持. 试试 `--model_name_or_path baichuan-inc/Baichuan-7B` 和 `--lora_target W_pack` 参数来使用 Baichuan-7B 模型.
|
|||
|
|
|||
|
[23/06/03] 现在我们支持了量化训练和推理 (也叫作 **[QLoRA](https://github.com/artidoro/qlora)**). 试试 `--quantization_bit 4/8` 参数来处理量化模型. (实验性功能)
|
|||
|
|
|||
|
[23/05/31] 现在我们在该仓库中提供了对于 **BLOOM & BLOOMZ** 模型的训练支持. 试试 `--model_name_or_path bigscience/bloomz-7b1-mt` 和 `--lora_target query_key_value` 参数来使用 BLOOMZ 模型.
|
|||
|
|
|||
|
## 支持的模型
|
|||
|
|
|||
|
- [LLaMA](https://github.com/facebookresearch/llama) (7B/13B/33B/65B)
|
|||
|
- [LLaMA-2](https://huggingface.co/meta-llama) (7B/13B/70B)
|
|||
|
- [BLOOM](https://huggingface.co/bigscience/bloom) & [BLOOMZ](https://huggingface.co/bigscience/bloomz) (560M/1.1B/1.7B/3B/7.1B/176B)
|
|||
|
- [Falcon](https://huggingface.co/tiiuae/falcon-7b) (7B/40B)
|
|||
|
- [Baichuan](https://huggingface.co/baichuan-inc/baichuan-7B) (7B/13B)
|
|||
|
- [InternLM](https://github.com/InternLM/InternLM) (7B)
|
|||
|
|
|||
|
## 支持的训练方法
|
|||
|
|
|||
|
- [(Continually) pre-training](https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf)
|
|||
|
- Full-parameter tuning
|
|||
|
- Partial-parameter tuning
|
|||
|
- [LoRA](https://arxiv.org/abs/2106.09685)
|
|||
|
- [QLoRA](https://arxiv.org/abs/2305.14314)
|
|||
|
- [Supervised fine-tuning](https://arxiv.org/abs/2109.01652)
|
|||
|
- Full-parameter tuning
|
|||
|
- Partial-parameter tuning
|
|||
|
- [LoRA](https://arxiv.org/abs/2106.09685)
|
|||
|
- [QLoRA](https://arxiv.org/abs/2305.14314)
|
|||
|
- [RLHF](https://arxiv.org/abs/2203.02155)
|
|||
|
- [LoRA](https://arxiv.org/abs/2106.09685)
|
|||
|
- [QLoRA](https://arxiv.org/abs/2305.14314)
|
|||
|
|
|||
|
## 提供的训练集
|
|||
|
|
|||
|
- 对预训练:
|
|||
|
- [Wiki Demo (en)](data/wiki_demo.txt)
|
|||
|
- 对监督微调:
|
|||
|
- [Stanford Alpaca (en)](https://github.com/tatsu-lab/stanford_alpaca)
|
|||
|
- [Stanford Alpaca (zh)](https://github.com/ymcui/Chinese-LLaMA-Alpaca)
|
|||
|
- [GPT-4 Generated Data (en&zh)](https://github.com/Instruction-Tuning-with-GPT-4/GPT-4-LLM)
|
|||
|
- [Open Assistant (multilingual)](https://huggingface.co/datasets/OpenAssistant/oasst1)
|
|||
|
- [Self-cognition (zh)](data/self_cognition.json)
|
|||
|
- [ShareGPT (zh)](https://huggingface.co/datasets/QingyiSi/Alpaca-CoT/tree/main/Chinese-instruction-collection)
|
|||
|
- [RefGPT (zh)](https://github.com/sufengniu/RefGPT)
|
|||
|
- [Guanaco Dataset (multilingual)](https://huggingface.co/datasets/JosephusCheung/GuanacoDataset)
|
|||
|
- [BELLE 2M (zh)](https://huggingface.co/datasets/BelleGroup/train_2M_CN)
|
|||
|
- [BELLE 1M (zh)](https://huggingface.co/datasets/BelleGroup/train_1M_CN)
|
|||
|
- [BELLE 0.5M (zh)](https://huggingface.co/datasets/BelleGroup/train_0.5M_CN)
|
|||
|
- [BELLE Dialogue 0.4M (zh)](https://huggingface.co/datasets/BelleGroup/generated_chat_0.4M)
|
|||
|
- [BELLE School Math 0.25M (zh)](https://huggingface.co/datasets/BelleGroup/school_math_0.25M)
|
|||
|
- [BELLE Multiturn Chat 0.8M (zh)](https://huggingface.co/datasets/BelleGroup/multiturn_chat_0.8M)
|
|||
|
- [Firefly 1.1M (zh)](https://huggingface.co/datasets/YeungNLP/firefly-train-1.1M)
|
|||
|
- [CodeAlpaca 20k (en)](https://huggingface.co/datasets/sahil2801/CodeAlpaca-20k)
|
|||
|
- [Alpaca CoT (multilingual)](https://huggingface.co/datasets/QingyiSi/Alpaca-CoT)
|
|||
|
- [Web QA (zh)](https://huggingface.co/datasets/suolyer/webqa)
|
|||
|
- [UltraChat (en)](https://github.com/thunlp/UltraChat)
|
|||
|
- [WebNovel (zh)](https://huggingface.co/datasets/zxbsmk/webnovel_cn)
|
|||
|
- 对奖励模型:
|
|||
|
- [HH-RLHF (en)](https://huggingface.co/datasets/Anthropic/hh-rlhf)
|
|||
|
- [Open Assistant (multilingual)](https://huggingface.co/datasets/OpenAssistant/oasst1)
|
|||
|
- [GPT-4 Generated Data (en&zh)](https://github.com/Instruction-Tuning-with-GPT-4/GPT-4-LLM)
|
|||
|
|
|||
|
更多细节请查看 [data/README.md](data/README_zh.md).
|
|||
|
|
|||
|
部分数据集使用之前需要确认, 因此推荐使用下面的命令登录您的 Huggingface 账户.
|
|||
|
|
|||
|
```bash
|
|||
|
pip install --upgrade huggingface_hub
|
|||
|
huggingface-cli login
|
|||
|
```
|
|||
|
|
|||
|
## 软件依赖
|
|||
|
|
|||
|
- Python 3.8+ 和 PyTorch 1.13.1+
|
|||
|
- 🤗Transformers, Datasets, Accelerate, PEFT 和 TRL
|
|||
|
- jieba, rouge-chinese 和 nltk (用于评估)
|
|||
|
- gradio 和 matplotlib (用于网页端交互)
|
|||
|
- uvicorn, fastapi 和 sse-starlette (用于 API)
|
|||
|
|
|||
|
以及 **强有力的 GPUs**!
|
|||
|
|
|||
|
如果要在 Windows 平台上开启量化 LoRA (QLoRA) , 需要安装预编译的 `bitsandbytes` 库, 支持 CUDA 11.1 到 12.1.
|
|||
|
|
|||
|
```bash
|
|||
|
pip install https://github.com/jllllll/bitsandbytes-windows-webui/releases/download/wheels/bitsandbytes-0.39.1-py3-none-win_amd64.whl
|
|||
|
```
|
|||
|
|
|||
|
## 起步
|
|||
|
|
|||
|
### 准备数据 (可选)
|
|||
|
|
|||
|
关于数据集文件的格式,请参考 `data/example_dataset` 文件夹的内容. 构建自定义数据集时, 既可以使用单个 `.json` 文件, 也可以使用一个[数据加载脚本](https://huggingface.co/docs/datasets/dataset_script)和多个文件.
|
|||
|
|
|||
|
注意:使用自定义数据集时,请更新 `data/dataset_info.json` 文件,该文件的格式请参考 `data/README.md`.
|
|||
|
|
|||
|
### 环境安装 (可选)
|
|||
|
|
|||
|
```bash
|
|||
|
git clone https://github.com/hiyouga/LLaMA-Efficient-Tuning.git
|
|||
|
conda create -n llama_etuning python=3.10
|
|||
|
conda activate llama_etuning
|
|||
|
cd LLaMA-Efficient-Tuning
|
|||
|
pip install -r requirements.txt
|
|||
|
```
|
|||
|
|
|||
|
### All-in-one Web UI
|
|||
|
|
|||
|
```bash
|
|||
|
python src/train_web.py
|
|||
|
```
|
|||
|
|
|||
|
### (Continually) Pre-Training
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage pt \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_train \
|
|||
|
--dataset wiki_demo \
|
|||
|
--finetuning_type lora \
|
|||
|
--output_dir path_to_pt_checkpoint \
|
|||
|
--overwrite_cache \
|
|||
|
--per_device_train_batch_size 4 \
|
|||
|
--gradient_accumulation_steps 4 \
|
|||
|
--lr_scheduler_type cosine \
|
|||
|
--logging_steps 10 \
|
|||
|
--save_steps 1000 \
|
|||
|
--learning_rate 5e-5 \
|
|||
|
--num_train_epochs 3.0 \
|
|||
|
--plot_loss \
|
|||
|
--fp16
|
|||
|
```
|
|||
|
|
|||
|
### 监督微调
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage sft \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_train \
|
|||
|
--dataset alpaca_gpt4_en \
|
|||
|
--finetuning_type lora \
|
|||
|
--output_dir path_to_sft_checkpoint \
|
|||
|
--overwrite_cache \
|
|||
|
--per_device_train_batch_size 4 \
|
|||
|
--gradient_accumulation_steps 4 \
|
|||
|
--lr_scheduler_type cosine \
|
|||
|
--logging_steps 10 \
|
|||
|
--save_steps 1000 \
|
|||
|
--learning_rate 5e-5 \
|
|||
|
--num_train_epochs 3.0 \
|
|||
|
--plot_loss \
|
|||
|
--fp16
|
|||
|
```
|
|||
|
|
|||
|
### 奖励模型训练
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage rm \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_train \
|
|||
|
--dataset comparison_gpt4_en \
|
|||
|
--finetuning_type lora \
|
|||
|
--output_dir path_to_rm_checkpoint \
|
|||
|
--per_device_train_batch_size 4 \
|
|||
|
--gradient_accumulation_steps 4 \
|
|||
|
--lr_scheduler_type cosine \
|
|||
|
--logging_steps 10 \
|
|||
|
--save_steps 1000 \
|
|||
|
--learning_rate 1e-5 \
|
|||
|
--num_train_epochs 1.0 \
|
|||
|
--plot_loss \
|
|||
|
--fp16
|
|||
|
```
|
|||
|
|
|||
|
### PPO Training (RLHF)
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage ppo \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_train \
|
|||
|
--dataset alpaca_gpt4_en \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_sft_checkpoint \
|
|||
|
--reward_model path_to_rm_checkpoint \
|
|||
|
--output_dir path_to_ppo_checkpoint \
|
|||
|
--per_device_train_batch_size 2 \
|
|||
|
--gradient_accumulation_steps 4 \
|
|||
|
--lr_scheduler_type cosine \
|
|||
|
--logging_steps 10 \
|
|||
|
--save_steps 1000 \
|
|||
|
--learning_rate 1e-5 \
|
|||
|
--num_train_epochs 1.0 \
|
|||
|
--resume_lora_training False \
|
|||
|
--plot_loss
|
|||
|
```
|
|||
|
|
|||
|
### 分布式微调
|
|||
|
|
|||
|
```bash
|
|||
|
accelerate config # configure the environment
|
|||
|
accelerate launch src/train_bash.py # arguments (same as above)
|
|||
|
```
|
|||
|
|
|||
|
<details><summary>使用 DeepSpeed ZeRO-2 全量微调的配置示例</summary>
|
|||
|
|
|||
|
```yaml
|
|||
|
compute_environment: LOCAL_MACHINE
|
|||
|
deepspeed_config:
|
|||
|
gradient_accumulation_steps: 4
|
|||
|
gradient_clipping: 0.5
|
|||
|
offload_optimizer_device: none
|
|||
|
offload_param_device: none
|
|||
|
zero3_init_flag: false
|
|||
|
zero_stage: 2
|
|||
|
distributed_type: DEEPSPEED
|
|||
|
downcast_bf16: 'no'
|
|||
|
machine_rank: 0
|
|||
|
main_training_function: main
|
|||
|
mixed_precision: fp16
|
|||
|
num_machines: 1
|
|||
|
num_processes: 4
|
|||
|
rdzv_backend: static
|
|||
|
same_network: true
|
|||
|
tpu_env: []
|
|||
|
tpu_use_cluster: false
|
|||
|
tpu_use_sudo: false
|
|||
|
use_cpu: false
|
|||
|
```
|
|||
|
|
|||
|
</details>
|
|||
|
|
|||
|
### 指标评估(BLEU分数和汉语ROUGE分数)
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage sft \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_eval \
|
|||
|
--dataset alpaca_gpt4_en \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint \
|
|||
|
--output_dir path_to_eval_result \
|
|||
|
--per_device_eval_batch_size 8 \
|
|||
|
--max_samples 100 \
|
|||
|
--predict_with_generate
|
|||
|
```
|
|||
|
|
|||
|
我们建议在 4/8-bit 评估中使用 `--per_device_eval_batch_size=1` 和 `--max_target_length 128`.
|
|||
|
|
|||
|
### 模型预测
|
|||
|
|
|||
|
```bash
|
|||
|
CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
|
--stage sft \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--do_predict \
|
|||
|
--dataset alpaca_gpt4_en \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint \
|
|||
|
--output_dir path_to_predict_result \
|
|||
|
--per_device_eval_batch_size 8 \
|
|||
|
--max_samples 100 \
|
|||
|
--predict_with_generate
|
|||
|
```
|
|||
|
|
|||
|
如果需要预测的样本没有标签,请首先在 `response` 列中填入一些占位符,以免样本在预处理阶段被丢弃。
|
|||
|
|
|||
|
### API 服务
|
|||
|
```bash
|
|||
|
python src/api_demo.py \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint
|
|||
|
```
|
|||
|
|
|||
|
访问 `http://localhost:8000/docs` 获取 API 文档.
|
|||
|
|
|||
|
### 命令行测试
|
|||
|
|
|||
|
```bash
|
|||
|
python src/cli_demo.py \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint
|
|||
|
```
|
|||
|
|
|||
|
### 浏览器测试
|
|||
|
|
|||
|
```bash
|
|||
|
python src/web_demo.py \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint
|
|||
|
```
|
|||
|
|
|||
|
### 导出微调模型
|
|||
|
|
|||
|
```bash
|
|||
|
python src/export_model.py \
|
|||
|
--model_name_or_path path_to_your_model \
|
|||
|
--finetuning_type lora \
|
|||
|
--checkpoint_dir path_to_checkpoint \
|
|||
|
--output_dir path_to_export
|
|||
|
```
|
|||
|
|
|||
|
## 协议
|
|||
|
|
|||
|
本仓库采用 [Apache-2.0](LICENSE) 协议开源.
|
|||
|
|
|||
|
请遵循模型许可证使用相应的模型权重:
|
|||
|
|
|||
|
- [LLaMA](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md)
|
|||
|
- [LLaMA-2](https://ai.meta.com/llama/license/)
|
|||
|
- [BLOOM](https://huggingface.co/spaces/bigscience/license)
|
|||
|
- [Falcon](LICENSE)
|
|||
|
- [baichuan](https://huggingface.co/baichuan-inc/baichuan-7B/resolve/main/baichuan-7B%20%E6%A8%A1%E5%9E%8B%E8%AE%B8%E5%8F%AF%E5%8D%8F%E8%AE%AE.pdf)
|
|||
|
- [InternLM](https://github.com/InternLM/InternLM#open-source-license)
|
|||
|
|
|||
|
## 引用
|
|||
|
|
|||
|
如果您觉得此项目有帮助,请考虑以下列格式引用:
|
|||
|
|
|||
|
```bibtex
|
|||
|
@Misc{llama-efficient-tuning,
|
|||
|
title = {LLaMA Efficient Tuning},
|
|||
|
author = {hiyouga},
|
|||
|
howpublished = {\url{https://github.com/hiyouga/LLaMA-Efficient-Tuning}},
|
|||
|
year = {2023}
|
|||
|
}
|
|||
|
```
|
|||
|
|
|||
|
## 致谢
|
|||
|
|
|||
|
该 repo 是 [ChatGLM-Efficient-Tuning](https://github.com/hiyouga/ChatGLM-Efficient-Tuning) 的兄弟产品. 它们都拥有在大规模语言模型上实现高效调优的相似的代码结构。
|
|||
|
|
|||
|
## Star History
|
|||
|
|
|||
|
![Star History Chart](https://api.star-history.com/svg?repos=hiyouga/LLaMA-Efficient-Tuning&type=Date)
|