ClusterFit

Last updated on Feb 27, 2021

ClusterFit ResNet-50 (ImageNet-1K, 16K RotNet clusters)

ClusterFit ResNet-50 (ImageNet-1K, 16K RotNet clusters) achieves 81.5% Top 1 Accuracy on ImageNet


Parameters 26 Million
FLOPs 4 Billion
File Size 429.88 MB
Training Data ImageNet
Training Resources 128 GPUs
Training Time

Training Techniques Rotnet, ClusterFit, Weight Decay, SGD with Momentum
Architecture 1x1 Convolution, Bottleneck Residual Block, Batch Normalization, Convolution, Global Average Pooling, Residual Block, Residual Connection, ReLU, Max Pooling, Softmax
ID rn50_in1k_clusterfit
Epochs 105
Layers 50
Classes 1000
Momentum 0.9
Weight Decay 0.0001
Width Multiplier 1
SHOW MORE
SHOW LESS
README.md

Summary

ClusterFit is a self-supervision approach for learning image representations. Given a dataset, we (a) cluster its features extracted from a pre-trained network using k-means and (b) re-train a new network from scratch on this dataset using cluster assignments as pseudo-labels.

How do I train this model?

Get started with VISSL by trying one of the Colab tutorial notebooks.

Citation

@article{DBLP:journals/corr/abs-1912-03330,
  author    = {Xueting Yan and
               Ishan Misra and
               Abhinav Gupta and
               Deepti Ghadiyaram and
               Dhruv Mahajan},
  title     = {ClusterFit: Improving Generalization of Visual Representations},
  journal   = {CoRR},
  volume    = {abs/1912.03330},
  year      = {2019},
  url       = {http://arxiv.org/abs/1912.03330},
  archivePrefix = {arXiv},
  eprint    = {1912.03330},
  timestamp = {Mon, 28 Sep 2020 08:19:37 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-1912-03330.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}
@misc{goyal2021vissl,
  author =       {Priya Goyal and Benjamin Lefaudeux and Mannat Singh and Jeremy Reizenstein and Vinicius Reis and 
                  Min Xu and and Matthew Leavitt and Mathilde Caron and Piotr Bojanowski and Armand Joulin and 
                  Ishan Misra},
  title =        {VISSL},
  howpublished = {\url{https://github.com/facebookresearch/vissl}},
  year =         {2021}
}

Results

Image Classification on ImageNet

Image Classification
BENCHMARK MODEL METRIC NAME METRIC VALUE GLOBAL RANK
ImageNet ClusterFit ResNet-50 (ImageNet-1K, 16K RotNet clusters) Top 1 Accuracy 53.63% # 309