ResNet Supervised

Last updated on

ResNet-101 (supervised, torchvision)

Parameters 45 Million
FLOPs 16 Billion
File Size 170.45 MB
Training Data ImageNet
Training Resources 8x NVIDIA V100 GPUs
Training Time

Training Techniques 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 rn101_torchvision
LR 0.1
Epochs 90
LR Gamma 0.1
Momentum 0.9
Batch Size 256
LR Step Size 30
Weight Decay 0.0001
SHOW MORE
SHOW LESS
ResNet-50 (supervised, caffe2)

Parameters 26 Million
FLOPs 4 Billion
File Size 89.96 MB
Training Data ImageNet
Training Resources 8x NVIDIA V100 GPUs
Training Time

Training Techniques 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_caffe2_in1k
Layers 50
Batch Size 256
Width Multiplier 1
SHOW MORE
SHOW LESS
ResNet-50 (supervised, caffe2, places205)

Parameters 26 Million
FLOPs 4 Billion
File Size 89.96 MB
Training Data Places205
Training Resources 8x NVIDIA V100 GPUs
Training Time

Training Techniques 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_caffe2_places205
Width Multiplier 1
SHOW MORE
SHOW LESS
ResNet-50 (supervised, torchvision)

Parameters 26 Million
FLOPs 4 Billion
File Size 97.75 MB
Training Data ImageNet
Training Resources 8x NVIDIA V100 GPUs
Training Time

Training Techniques 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_torchvision
LR 0.1
Epochs 105
Layers 50
LR Gamma 0.1
Momentum 0.9
Batch Size 256
LR Step Size 30
Weight Decay 0.0001
Width Multiplier 1
SHOW MORE
SHOW LESS
ResNet-50 (supervised, vissl)

Parameters 26 Million
FLOPs 4 Billion
File Size 195.40 MB
Training Data ImageNet
Training Resources 8x NVIDIA V100 GPUs
Training Time

Training Techniques 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_vissl
LR 0.1
Epochs 105
Layers 50
Momentum 0.9
Batch Size 256
Weight Decay 0.0001
Width Multiplier 1
SHOW MORE
SHOW LESS
README.md

Summary

ResNet Supervised is a model collection of residual networks trained with regular supervised learning. Residual Networks, or ResNets, learn residual functions with reference to the layer inputs, instead of learning unreferenced functions. Instead of hoping each few stacked layers directly fit a desired underlying mapping, residual nets let these layers fit a residual mapping. They stack residual blocks ontop of each other to form network: e.g. a ResNet-50 has fifty layers using these blocks.

You can use these models as reference points with which to compare models trained with self-supervised or semi-supervised learning.

How do I train this model?

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

Citation

@article{DBLP:journals/corr/HeZRS15,
  author    = {Kaiming He and
               Xiangyu Zhang and
               Shaoqing Ren and
               Jian Sun},
  title     = {Deep Residual Learning for Image Recognition},
  journal   = {CoRR},
  volume    = {abs/1512.03385},
  year      = {2015},
  url       = {http://arxiv.org/abs/1512.03385},
  archivePrefix = {arXiv},
  eprint    = {1512.03385},
  timestamp = {Wed, 17 Apr 2019 17:23:45 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/HeZRS15.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 on ImageNet
MODEL TOP 1 ACCURACY
ResNet-101 (supervised, torchvision) 77.21%
ResNet-50 (supervised, torchvision) 76.1%
ResNet-50 (supervised, caffe2) 75.88%
ResNet-50 (supervised, vissl) 75.45%
ResNet-50 (supervised, caffe2, places205) 58.49%