ChatGLM Efficient Tuning

Fine-tuning 🤖ChatGLM-6B model with 🤗PEFT.

👋 Join our WeChat.

[ English | 中文 ]

Changelog

[23/04/29] Now we support training ChatGLM with Reinforcement Learning with Human Feedback (RLHF) ! We provide several examples to run RLHF training, please refer to the examples folder for details. (experimental feature)

[23/04/20] Our repo achieved 100 stars within 12 days! Congratulations!

[23/04/19] Now we support merging the weights of fine-tuned models trained by LoRA! Try --checkpoint_dir checkpoint1,checkpoint2 argument for continually fine-tuning the models.

[23/04/18] Now we support training the quantized models using three fine-tuning methods! Try quantization_bit argument for training the model in 4/8 bits.

[23/04/12] Now we support training from checkpoints! Use --checkpoint_dir argument to specify the checkpoint model to fine-tune from.

[23/04/11] Now we support training with combined datasets! Try --dataset dataset1,dataset2 argument for training with multiple datasets.

Datasets

Our script now supports the following datasets:

Please refer to data/README.md for details.

Some datasets require confirmation before using them, so we recommend logging in with your HuggingFace account using these commands.

pip install --upgrade huggingface_hub
huggingface-cli login

Fine-Tuning Methods

Our script now supports the following fine-tuning methods:

LoRA
- Fine-tuning the low-rank adapters of the model.
P-Tuning V2
- Fine-tuning the prefix encoder of the model.
Freeze
- Fine-tuning the MLPs in the last n blocks of the model.

Requirement

Python 3.8+ and PyTorch 1.13.1
🤗Transformers, Datasets, Accelerate, PEFT and TRL
protobuf, cpm_kernels and sentencepiece
jieba, rouge_chinese and nltk (used at evaluation)
gradio and mdtex2html (used in web_demo.py)

And powerful GPUs!

Getting Started

Data Preparation (optional)

Please refer to data/example_dataset for checking the details about the format of dataset files. You can either use a single .json file or a dataset loading script with multiple files to create a custom dataset.

Note: please update data/dataset_info.json to use your custom dataset. About the format of this file, please refer to data/README.md.

Dependence Installation (optional)

git clone https://github.com/hiyouga/ChatGLM-Efficient-Tuning.git
conda create -n chatglm_etuning python=3.10
conda activate chatglm_etuning
cd ChatGLM-Efficient-Tuning
pip install -r requirements.txt

If you want to enable LoRA or Freeze quantization on Windows, you will be required to install a pre-built version of bitsandbytes library, which supports CUDA 11.6 or 11.7.

pip install https://github.com/acpopescu/bitsandbytes/releases/download/v0.37.2-win.1/bitsandbytes-0.37.2-py3-none-any.whl

Fine-tuning with a Single GPU

CUDA_VISIBLE_DEVICES=0 python src/finetune.py \
    --do_train \
    --dataset alpaca_gpt4_zh \
    --finetuning_type lora \
    --output_dir path_to_checkpoint \
    --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 1.0 \
    --fp16

Please refer to our Wiki about the details of the arguments.

Distributed Fine-tuning with Multiple GPUs

accelerate config # configure the environment
accelerate launch src/finetune.py # arguments (same as above)

Note: if you are using LoRA method at fine-tuning, please provide --ddp_find_unused_parameters False argument to avoid the runtime error.

Training Reward Model

CUDA_VISIBLE_DEVICES=0 python src/train_rm.py \
    --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 5e-5 \
    --num_train_epochs 1.0 \
    --fp16

Training with RLHF

CUDA_VISIBLE_DEVICES=0 python src/train_ppo.py \
    --do_train \
    --dataset alpaca_gpt4_en \
    --finetuning_type lora \
    --reward_model path_to_rm_checkpoint \
    --output_dir path_to_ppo_checkpoint \
    --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 1.0 \
    --fp16

Evaluation (BLEU and ROUGE_CHINESE)

CUDA_VISIBLE_DEVICES=0 python src/finetune.py \
    --do_eval \
    --dataset alpaca_gpt4_zh \
    --checkpoint_dir path_to_checkpoint \
    --output_dir path_to_eval_result \
    --per_device_eval_batch_size 8 \
    --max_samples 50 \
    --predict_with_generate

Predict

CUDA_VISIBLE_DEVICES=0 python src/finetune.py \
    --do_predict \
    --dataset alpaca_gpt4_zh \
    --checkpoint_dir path_to_checkpoint \
    --output_dir path_to_predict_result \
    --per_device_eval_batch_size 8 \
    --max_samples 50 \
    --predict_with_generate

Inference

CUDA_VISIBLE_DEVICES=0 python src/infer.py \
    --checkpoint_dir path_to_checkpoint

Web Demo

CUDA_VISIBLE_DEVICES=0 python src/web_demo.py \
    --checkpoint_dir path_to_checkpoint

Deploy the Fine-tuned Model

import sys
sys.path.append("src")
from src import load_pretrained, ModelArguments
model_args = ModelArguments(checkpoint_dir=path_to_checkpoint)
model, tokenizer = load_pretrained(model_args)
model = model.cuda()
model.eval()
# model.generate, model.chat()...

Hardware Requirements

Fine-tune method	Batch size	Mode	GRAM	Speed
LoRA (r=8)	16	FP16	28GB	8ex/s
LoRA (r=8)	8	FP16	24GB	8ex/s
LoRA (r=8)	4	FP16	20GB	8ex/s
LoRA (r=8)	4	INT8	10GB	8ex/s
P-Tuning (p=16)	4	FP16	20GB	8ex/s
P-Tuning (p=16)	4	INT8	16GB	8ex/s
P-Tuning (p=16)	4	INT4	12GB	8ex/s
Freeze (l=3)	4	FP16	24GB	8ex/s
Freeze (l=3)	4	INT8	12GB	8ex/s

RM method	Batch size	Mode	GRAM	Speed
LoRA (r=8) + rm	4	FP16	22GB	-
LoRA (r=8) + rm	1	INT8	11GB	-

RLHF method	Batch size	Mode	GRAM	Speed
LoRA (r=8) + ppo	4	FP16	23GB	-
LoRA (r=8) + ppo	1	INT8	12GB	-

Note: r is the lora rank, p is the number of prefix tokens, l is the number of trainable layers, ex/s is the examples per second at training. The gradient_accumulation_steps is set to 1. All are evaluated on a single Tesla V100 (32G) GPU, they are approximated values and may vary in different GPUs.

Fine-tuning ChatGLM: A Case

Training Results

We use the whole alpaca_gpt4_zh dataset to fine-tune the ChatGLM model with LoRA (r=8) for one epoch, using the default hyper-parameters. The loss curve during training is presented below.

Evaluation Results

We select 100 instances in the alpaca_gpt4_zh dataset to evaluate the fine-tuned ChatGLM model and compute the BLEU and ROUGE scores. The results are presented below.

Score	Original	FZ (l=2)	PT (p=16)	LoRA (r=8)
BLEU-4	15.75	16.85	16.06	17.01 (+1.26)
Rouge-1	34.51	36.62	34.80	36.77 (+2.26)
Rouge-2	15.11	17.04	15.32	16.83 (+1.72)
Rouge-l	26.18	28.17	26.35	28.86 (+2.68)
Params (%)	/	4.35%	0.06%	0.06%

FZ: freeze tuning, PT: P-Tuning V2 (we use pre_seq_len=16 for fair comparison with LoRA), Params: the percentange of trainable parameters.

Compared with Existing Implementations

THUDM/ChatGLM-6B
- Official implementation of fine-tuning ChatGLM with P-Tuning v2 on the ADGEN dataset.
- Our fine-tuning script is largely depend on it. We further implement the LoRA tuning method. Additionally, we dynamically pad the inputs to the longest sequence in the batch instead of the maximum length, to accelerate the fine-tuning.
mymusise/ChatGLM-Tuning
- An unoffical implementation of fine-tuning ChatGLM with LoRA on the Stanford Alpaca dataset.
- We borrowed some ideas from it. Our fine-tuning script integrates the data pre-processing part into the training procedure, so we need not generate a pre-processed dataset before training.
ssbuild/chatglm_finetuning
- An unofficial implementation of fine-tuning ChatGLM with several PEFT methods on the Stanford Alpaca dataset.
- Our fine-tuning script is implemented purely with Huggingface transformers and is independent of the deep_training framework.
lich99/ChatGLM-finetune-LoRA
- An unofficial implementation of fine-tuning ChatGLM with LoRA on the Stanford Alpaca dataset.
- We use the Huggingface PEFT to provide the state-of-the-art PEFT methods.
liucongg/ChatGLM-Finetuning
- An unofficial implementation of fine-tuning ChatGLM with several methods including Freeze, LoRA and P-Tuning on the industrial dataset.
- We are aim to incorporate more instruction-following datasets for fine-tuning the ChatGLM model.
yanqiangmiffy/InstructGLM
- An unofficial implementation of fine-tuning ChatGLM that explores the ChatGLM's ability on the instruction-following datasets.
- Our fine-tuning script integrates the data pre-processing part in to the training procedure.

TODO

License

This repository is licensed under the Apache-2.0 License. Please follow the Model License to use ChatGLM-6B model.

Citation

If this work is helpful, please cite as:

@Misc{chatglm-efficient-tuning,
  title = {ChatGLM Efficient Tuning},
  author = {hiyouga},
  howpublished = {\url{https://github.com/hiyouga/ChatGLM-Efficient-Tuning}},
  year = {2023}
}

Acknowledgement

This repo benefits from ChatGLM-6B, ChatGLM-Tuning and yuanzhoulvpi2017/zero_nlp. Thanks for their wonderful works.

jiluojiluo / ChatGLM-Efficient-Tuning