MoSA: Mixture of Sparse Adapters for Visual Efficient Tuning
- URL: http://arxiv.org/abs/2312.02923v2
- Date: Sat, 23 Mar 2024 07:00:15 GMT
- Title: MoSA: Mixture of Sparse Adapters for Visual Efficient Tuning
- Authors: Qizhe Zhang, Bocheng Zou, Ruichuan An, Jiaming Liu, Shanghang Zhang,
- Abstract summary: Mixture of Sparse Adapters, or MoSA, is a novel Adapter Tuning method.
MoSA can achieve significantly better performance than standard without any additional computational storage overhead.
MoSA consistently outperforms other Adapter Tuning methods as well as other baselines by a large margin.
- Score: 20.68925288222065
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: With the rapid growth in the scale of pre-trained foundation models, parameter-efficient fine-tuning techniques have gained significant attention, among which Adapter Tuning is the most widely used. Despite achieving efficiency, it still underperforms full fine-tuning, and the performance improves at the cost of an increase in parameters. Recent efforts have either focused on training multiple adapter experts to increase model capacity or on pruning adapters to achieve parameter efficiency. However, both approaches introduce more parameters compared to the original adapter, hence are not computationally efficient. Motivated by this, we propose Mixture of Sparse Adapters, or MoSA, as a novel Adapter Tuning method to fully unleash the potential of each parameter in the adapter. We first split the standard adapter into multiple non-overlapping modules, then stochastically activate them for sparse training, and finally merge them to form a complete adapter after tuning. In this way, MoSA can achieve significantly better performance than standard adapters without any additional computational or storage overhead. Furthermore, we propose a hierarchical sparse strategy to better leverage limited training data. Extensive experiments on a series of 27 visual tasks demonstrate that MoSA consistently outperforms other Adapter Tuning methods as well as other baselines by a large margin. Furthermore, MoSA brings consistent improvements across various model scales, architectures, and different PEFT methods. Code will be released.
Related papers
- Hadamard Adapter: An Extreme Parameter-Efficient Adapter Tuning Method for Pre-trained Language Models [108.08773541490191]
Pre-trained Language models (PLMs) have a huge amount of parameters, fine-tuning them is often expensive and time consuming.
It is necessary to adopt a parameter-efficient approach to reduce parameters of PLMs in fine-tuning without compromising their performance in downstream tasks.
In this paper, we design a novel adapter which only acts on self-attention outputs in PLMs.
arXiv Detail & Related papers (2024-07-04T18:21:28Z) - MerA: Merging Pretrained Adapters For Few-Shot Learning [71.44422347502409]
We propose textbftextttMerging Pretrained Adapters (MerA) that efficiently incorporates pretrained adapters to a single model through model fusion.
Experiments on two PLMs demonstrate that MerA substantial improvements compared to both single adapters and AdapterFusion.
arXiv Detail & Related papers (2023-08-30T12:10:17Z) - A Comprehensive Analysis of Adapter Efficiency [20.63580880344425]
We show that for Natural Language Understanding (NLU) tasks, the parameter efficiency in adapters does not translate to efficiency gains compared to full fine-tuning of models.
We recommend that for moderately sized models for NLU tasks, practitioners should rely on full fine-tuning or multi-task training rather than using adapters.
arXiv Detail & Related papers (2023-05-12T14:05:45Z) - SparseAdapter: An Easy Approach for Improving the Parameter-Efficiency
of Adapters [96.52807311742198]
We re-examine the parameter-efficiency of Adapters through the lens of network pruning.
We find that SparseAdapter can achieve comparable or better performance than standard Adapters when the sparse ratio reaches up to 80%.
arXiv Detail & Related papers (2022-10-09T15:28:48Z) - AdaMix: Mixture-of-Adapter for Parameter-efficient Tuning of Large
Language Models [119.7093605087114]
Fine-tuning large-scale pre-trained language models to downstream tasks require updating hundreds of millions of parameters.
This not only increases the serving cost to store a large copy of the model weights for every task, but also exhibits instability during few-shot task adaptation.
We introduce a new mechanism to improve adapter capacity without increasing parameters or computational cost by two key techniques.
arXiv Detail & Related papers (2022-05-24T23:41:22Z) - Adaptable Adapters [74.65986170056945]
State-of-the-art pretrained NLP models contain a hundred million to trillion parameters.
Adaptable adapters contain different activation functions for different layers and different input data.
We show that adaptable adapters achieve on-par performances with the standard adapter architecture.
arXiv Detail & Related papers (2022-05-03T14:59:27Z) - AdapterBias: Parameter-efficient Token-dependent Representation Shift
for Adapters in NLP Tasks [55.705355299065474]
Transformer-based pre-trained models with millions of parameters require large storage.
Recent approaches tackle this shortcoming by training adapters, but these approaches still require a relatively large number of parameters.
In this study, AdapterBias, a surprisingly simple yet effective adapter architecture, is proposed.
arXiv Detail & Related papers (2022-04-30T16:49:41Z) - On the Effectiveness of Adapter-based Tuning for Pretrained Language
Model Adaptation [36.37565646597464]
adapter-based tuning works by adding light-weight adapter modules to a pretrained language model (PrLM)
It adds only a few trainable parameters per new task, allowing a high degree of parameter sharing.
We demonstrate that adapter-based tuning outperforms fine-tuning on low-resource and cross-lingual tasks.
arXiv Detail & Related papers (2021-06-06T16:10:12Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.