How Well Do Sparse ImageNet Models Transfer?

By Eugenia Iofinova, Alexandra Peste, Mark Kurtz, Dan Alistarh. 2021

53 citations recorded by the source corpus.

Transfer learning is a classic paradigm by which models pretrained on large “upstream” datasets are adapted to yield good results on “downstream” specialized datasets. Generally, more accurate models on the “upstream” dataset tend to provide better transfer accuracy “downstream”. In this work, we perform an in-depth investigation of this phenomenon in the context of convolutional neural networks (CNNs) trained on the ImageNet dataset, which have been pruned-that is, compressed by sparsifiying their connections. We consider transfer using unstructured pruned models obtained by applying several state-of-the-art pruning methods, including magnitude-based, second-order, regrowth, lottery-ticket, and regularization approaches, in the context of twelve standard transfer tasks. In a nutshell, our study shows that sparse models can match or even outperform the transfer performance of dense models, even at high sparsities, and, while doing so, can lead to significant inference and even training speedups. At the same time, we observe and analyze significant differences in the behaviour of different pruning methods. The code is available at: https://github.com/IST-DASLab/sparse-imagenet-transfer.

Subfields: Artificial Intelligence, Computational Mechanics, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Elias Frantar
2. Dan Alistarh
3. Torsten Hoefler
4. Saleh Ashkboos
5. Peter Richtárik
6. Tim Dettmers
7. Sebastian U. Stich
8. Martin Jaggi

In Computational Mechanics

1. Dan Alistarh
2. Elias Frantar
3. Torsten Hoefler

In Computer Vision and Pattern Recognition

1. Saleh Ashkboos
2. Elias Frantar
3. Torsten Hoefler
4. Dan Alistarh
5. Tim Dettmers
6. Luke S. Zettlemoyer
7. Peter Richtárik

In Representation & Self-Supervised Learning

1. Torsten Hoefler
2. Dan Alistarh

Citations in the corpus

• GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers (2022, background): “We believe a thorough study of the impact of compression upon secondary measures, and in particular transferrability [15] or bias effects [2] is warranted, and may be rendered easier through our work.”
• A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations (2023, unknown): “In this section, we discuss how to choose different pruning methods and provide promising directions for future work.”