Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

UniPT: Universal Parallel Tuning for Transfer Learning with Efficient Parameter and Memory

About

Parameter-efficient transfer learning (PETL), i.e., fine-tuning a small portion of parameters, is an effective strategy for adapting pre-trained models to downstream domains. To further reduce the memory demand, recent PETL works focus on the more valuable memory-efficient characteristic. In this paper, we argue that the scalability, adaptability, and generalizability of state-of-the-art methods are hindered by structural dependency and pertinency on specific pre-trained backbones. To this end, we propose a new memory-efficient PETL strategy, Universal Parallel Tuning (UniPT), to mitigate these weaknesses. Specifically, we facilitate the transfer process via a lightweight and learnable parallel network, which consists of: 1) A parallel interaction module that decouples the sequential connections and processes the intermediate activations detachedly from the pre-trained network. 2) A confidence aggregation module that learns optimal strategies adaptively for integrating cross-layer features. We evaluate UniPT with different backbones (e.g., T5, VSE$\infty$, CLIP4Clip, Clip-ViL, and MDETR) on various vision-and-language and pure NLP tasks. Extensive ablations on 18 datasets have validated that UniPT can not only dramatically reduce memory consumption and outperform the best competitor, but also achieve competitive performance over other plain PETL methods with lower training memory overhead. Our code is publicly available at: https://github.com/Paranioar/UniPT.

Haiwen Diao, Bo Wan, Ying Zhang, Xu Jia, Huchuan Lu, Long Chen• 2023

Related benchmarks

TaskDatasetResultRank
Visual Question AnsweringVQA v2 (test-dev)
Overall Accuracy75.33
664
Text-to-Image RetrievalFlickr30K
R@167.7
460
Text-to-Video RetrievalMSR-VTT
Recall@138.9
313
Visual Question AnsweringGQA (test-dev)
Accuracy60.1
178
Visual GroundingRefCOCO+ (val)
Accuracy72.94
171
Visual GroundingRefCOCO+ (testB)
Accuracy64.49
169
Visual GroundingRefCOCO+ (testA)
Accuracy79.18
168
Video-to-Text retrievalMSR-VTT
Recall@139.3
157
Visual GroundingRefCOCO (testB)
Accuracy78.16
125
Visual GroundingRefCOCO (val)
Accuracy82.71
119
Showing 10 of 23 rows

Other info

Code

Follow for update