# How Well Do Sparse Imagenet Models Transfer? Source: [How Well Do Sparse Imagenet Models Transfer?]() Collection index: [llms.txt]() Data updated: 2025-12-24T17:12:06.013Z Authors: Eugenia Iofinova, Alexandra Peste, Mark Kurtz, Dan Alistarh Published: 2021-11-26T11:58:51.000Z ## Abstract 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 sparsifying their connections. We consider transfer using unstructured pruned models obtained by applying several state-of-the-art pruning methods, including magnitude-based, second-order, re-growth, 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. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on FGVC Suite](): Image Classification; FGVC; Average Accuracy: 88.54. - [Image Classification on VTAB 1k (test)](): Image Classification; VTAB 1k (test); Accuracy (Natural): 71.95. - [Image Classification on VTAB 1k](): Image Classification; VTAB 1K; Overall Mean Accuracy: 65.57. - [Visual Task Adaptation on VTAB 1K](): Visual Task Adaptation; VTAB 1K; Average Accuracy: not recorded. ## Original sources and implementation [Original paper]() [PDF]()