Conditional adapters: Parameter-efficient transfer learning with fast inference

T Lei, J Bai, S Brahma, J Ainslie… - Advances in …, 2023 - proceedings.neurips.cc
Abstract We propose Conditional Adapter (CoDA), a parameter-efficient transfer learning
method that also improves inference efficiency. CoDA generalizes beyond standard adapter …

Differentiable transportation pruning

Y Li, JC van Gemert, T Hoefler… - Proceedings of the …, 2023 - openaccess.thecvf.com
Deep learning algorithms are increasingly employed at the edge. However, edge devices
are resource constrained and thus require efficient deployment of deep neural networks …

Advancing dynamic sparse training by exploring optimization opportunities

J Ji, G Li, L Yin, M Qin, G Yuan, L Guo… - Forty-First International …, 2024 - openreview.net
Dynamic Sparse Training (DST) is an effective approach for addressing the substantial
training resource requirements posed by the ever-increasing size of the Deep Neural …

Is overfitting necessary for implicit video representation?

HM Choi, H Kang, D Oh - International Conference on …, 2023 - proceedings.mlr.press
Compact representation of multimedia signals using implicit neural representations (INRs)
has advanced significantly over the past few years, and recent works address their …

PETAH: Parameter Efficient Task Adaptation for Hybrid Transformers in a resource-limited Context

M Augustin, SS Sarwar, M Elhoushi, SQ Zhang… - arxiv preprint arxiv …, 2024 - arxiv.org
Following their success in natural language processing (NLP), there has been a shift
towards transformer models in computer vision. While transformers perform well and offer …

Mixed Sparsity Training: Achieving 4 FLOP Reduction for Transformer Pretraining

P Hu, S Li, L Huang - arxiv preprint arxiv:2408.11746, 2024 - arxiv.org
Large language models (LLMs) have made significant strides in complex tasks, yet their
widespread adoption is impeded by substantial computational demands. With hundreds of …