Model Zoo — MMSelfSup 1.0.0 documentation (original) (raw)

Shortcuts

All models and part of benchmark results are recorded below.

Benchmarks

ImageNet

ImageNet has multiple versions, but the most commonly used one is ILSVRC 2012. The classification results below are reported by linear evaluation or fine-tuning with pre-trained weights provided by various algorithms.

Algorithm Backbone Epoch Batch Size Results (Top-1 %) Links
Linear Eval Fine-tuning Pretrain Linear Eval Fine-tuning
Relative-Loc ResNet50 70 512 40.4 / config | model log config | model log /
Rotation-Pred ResNet50 70 128 47.0 / config | model log config | model log /
NPID ResNet50 200 256 58.3 / config | model log config | model log /
SimCLR ResNet50 200 256 62.7 / config | model log config | model log /
ResNet50 200 4096 66.9 / config | model log config | model log /
ResNet50 800 4096 69.2 / config | model log config | model log /
MoCo v2 ResNet50 200 256 67.5 / config | model log config | model log /
BYOL ResNet50 200 4096 71.8 / config | model log config | model log /
SwAV ResNet50 200 256 70.5 / config | model log config | model log /
DenseCL ResNet50 200 256 63.5 / config | model log config | model log /
SimSiam ResNet50 100 256 68.3 / config | model log config | model log /
ResNet50 200 256 69.8 / config | model log config | model log /
BarlowTwins ResNet50 300 2048 71.8 / config | model log config | model log /
MoCo v3 ResNet50 100 4096 69.6 / config | model log config | model log /
ResNet50 300 4096 72.8 / config | model log config | model log /
ResNet50 800 4096 74.4 / config | model log config | model log /
ViT-small 300 4096 73.6 / config | model log config | model log /
ViT-base 300 4096 76.9 83.0 config | model log config | model log config | model log
ViT-large 300 4096 / 83.7 config | model log / config | model log
MAE ViT-base 300 4096 60.8 82.8 config | model log config | model log config | model log
ViT-base 400 4096 62.5 83.3 config | model log config | model log config | model log
ViT-base 800 4096 65.1 83.3 config | model log config | model log config | model log
ViT-base 1600 4096 67.1 83.5 config | model log config | model log config | model log
ViT-large 400 4096 70.7 85.2 config | model log config | model log config | model log
ViT-large 800 4096 73.7 85.4 config | model log config | model log config | model log
ViT-large 1600 4096 75.5 85.7 config | model log config | model log config | model log
ViT-huge-FT-224 1600 4096 / 86.9 config | model log / config | model log
ViT-huge-FT-448 1600 4096 / 87.3 config | model log / config | model log
CAE ViT-base 300 2048 / 83.3 config | model log / config | model log
SimMIM Swin-base-FT192 100 2048 / 82.7 config | model log / config | model log
Swin-base-FT224 100 2048 / 83.5 config | model log / config | model log
Swin-base-FT224 800 2048 / 83.7 config | model log / config | model log
Swin-large-FT224 800 2048 / 84.8 config | model log / config | model log
MaskFeat ViT-base 300 2048 / 83.4 config | model log / config | model log
BEiT ViT-base 300 2048 / 83.1 config | model log / config | model log
MILAN ViT-base 400 4096 78.9 85.3 config | model log config | model log config | model log
BEiT v2 ViT-base 300 2048 / 85.0 config | model log / config | model log
EVA ViT-base 400 4096 69.0 83.7 config | model log config | model log config | model log
MixMIM MixMIM-Base 400 2048 / 84.6 config | model log / config | model log
PixMIM ViT-base 300 4096 63.3 83.1 config | model log config | model log config | model log
ViT-base 800 4096 67.5 83.5 config | model log config | model log config | model log