test-mass-forker-org-1 / adamix

This is the implementation of the paper AdaMix: Mixture-of-Adapter for Parameter-efficient Tuning of Large Language Models(https://arxiv.org/abs/2205.12410).

AdaMix (Mixture-of-Adapter)

This is the implementation of the paper AdaMix: Mixture-of-Adapter for Parameter-efficient Tuning of Large Language Models.

Quick links

• Overview
• GLUE Benchmark
  • BERT-base
  • RoBERTa-large
  • Averaging Weigths v.s. Random Routing
• Checkpoints
• Run the model
  • Quick start
  • Evaluate checkpoints
• Notes and Acknowledgments
• Contact Information
• Citation

Overview

Training

Inference

Adapting to the GLUE Benchmark

Our experiments on the GLUE benchmark are run on 16 NVIDIA Tesla V100 GPU. The results may vary due to different GPU models, drivers, CUDA SDK versions, floating-point precisions, and random seeds.

Main Results (BERT-base)

Main Results (RoBERTa-large)

Comparision between Averaging Weights and Random Single Adapter

Download AdaMix checkpoints

We release all copies of Adapter weights for users' Adapter aggregation study.

	Dataset	BERT base 110M	RoBERTa large 355M
	MNLI	8.5 MB	11.7 MB
	SST2	8.5 MB	11.7 MB
	MRPC	8.5 MB	11.7 MB
	CoLA	8.5 MB	11.7 MB
	QNLI	8.5 MB	11.7 MB
	QQP	8.5 MB	11.7 MB
	RTE	8.5 MB	11.7 MB
	STSB	8.5 MB	11.7 MB

Steps to reproduce our results

Create and activate conda env

conda env create -f environment.yml

Install the pre-requisites

pip install -e .

We also provide the shell scripts for bert-base and roberta-large.

Quick start

export num_gpus=1
export PYTHONHASHSEED=0
task_name=mnli
model=roberta-large
export output_dir="./models/${model}/${task_name}"
python -m torch.distributed.launch --nproc_per_node=$num_gpus \
examples/text-classification/run_glue.py \
--model_name_or_path $model \
--task_name $task_name \
--do_train \
--do_eval \
--max_seq_length 128 \
--per_device_train_batch_size 64 \
--per_device_eval_batch_size 32 \
--learning_rate 3e-4 \
--num_train_epochs 20 \
--output_dir $output_dir/model \
--overwrite_output_dir \
--logging_steps 1000 \
--logging_dir $output_dir/log \
--evaluation_strategy epoch \
--save_strategy epoch \
--warmup_ratio 0.06 \
--apply_expert_soup \
--adapter_size 16 \
--num_experts 4 \
--seed 0 \
--inference_level 3 \
--weight_decay 0.1 \
--sharing_up 1 \
--sharing_down 0 \
--use_consistency_loss 1

Most arguments are inherited from transformers and are easy to understand. We further explain some of the AdaMix's arguments:

• inference_level: There are two suggested modes

  • 1: Random Routing
  • 3: Averaging the weights of Adapters for routing (used in AdaMix)

• num_experts: Number of Adapters in AdaMix

• use_consistency_loss: Two modes.

  • 0: No consistency loss
  • 1: Use consistency loss

• sharing_up: There are two modes. (sharing_down is same)

  • 0: No weight sharing
  • 1: Sharing Project-up layer weights in Adapter

Evaluate the checkpoints

Create checkpoints directory and download checkpoints of corresponding tasks under the directory. Use MNLI as an example. Use your checkpoint path in expert_soup_path argument.

export num_gpus=1
export PYTHONHASHSEED=0
task_name=mnli
model=roberta-large
export output_dir="./models/${model}/${task_name}"
python -m torch.distributed.launch --nproc_per_node=$num_gpus \
examples/text-classification/run_glue.py \
--model_name_or_path $model \
--task_name $task_name \
--do_eval \
--expert_soup_path ./checkpoints/pytorch_model_${task_name}_expert_soup.bin \
--max_seq_length 128 \
--per_device_train_batch_size 64 \
--per_device_eval_batch_size 32 \
--learning_rate 3e-4 \
--num_train_epochs 20 \
--output_dir $output_dir/model \
--overwrite_output_dir \
--logging_steps 1000 \
--logging_dir $output_dir/log \
--evaluation_strategy epoch \
--save_strategy epoch \
--warmup_ratio 0.06 \
--apply_expert_soup \
--adapter_size 16 \
--num_experts 4 \
--seed 0 \
--inference_level 3 \
--weight_decay 0.1 \
--sharing_up 1 \
--sharing_down 0 \
--use_consistency_loss 1

Notes and Acknowledgments

The implementation is based on https://github.com/huggingface/transformers
We also used some code from: https://github.com/microsoft/LoRA

Contact Information

For personal communication related to this package, please contact Yaqing Wang ([email protected]), Subhabrata (Subho) Mukherjee ([email protected]) or Xiaodong Liu ([email protected]).

How do I cite AdaMix?

@article{wang2022adamix,
  title={AdaMix: Mixture-of-Adapter for Parameter-efficient Tuning of Large Language Models},
  author={Wang, Yaqing and Mukherjee, Subhabrata and Liu, Xiaodong and Gao, Jing and Awadallah, Ahmed Hassan and Gao, Jianfeng},
  journal={arXiv preprint arXiv:2205.12410},
  year={2022}
}