Awesome
Lottery Jackpots Exist in Pre-trained Models (Paper Link)
Requirements
- Python >= 3.7.4
- Pytorch >= 1.6.1
- Torchvision >= 0.4.1
Reproduce the Experiment Results
-
Download the pre-trained models from this link and place them in the
pre-trainfolder. -
Select a configuration file in
configsto reproduce the experiment results reported in the paper. For example, to find a lottery jackpot with 30 epochs for pruning 95% parameters of ResNet-32 on CIFAR-10, run:python cifar.py --config configs/resnet32_cifar10/90sparsity30epoch.yaml --gpus 0To find a lottery jackpot with 30 epochs for pruning 90% parameters of ResNet-50 on ImageNet, run:
python imagenet.py --config configs/resnet50_imagenet/90sparsity30epoch.yaml --gpus 0Note that the
data_pathin the yaml file should be changed to the data.
Evaluate Our Pruned Models
We provide configuration, training logs, and pruned models reported in the paper. They can be downloaded from the provided links in the following table:
| Model | Dataset | Sparsity | Epoch | Top-1 Acc. | Link |
|---|---|---|---|---|---|
| VGGNet-19 | CIFAR-10 | 90% | 30(S) | 93.88% | link |
| VGGNet-19 | CIFAR-10 | 90% | 160(S) | 93.94% | link |
| VGGNet-19 | CIFAR-10 | 95% | 30(S) | 93.49% | link |
| VGGNet-19 | CIFAR-10 | 95% | 160(S) | 93.74% | link |
| VGGNet-19 | CIFAR-100 | 90% | 30(S) | 72.59% | link |
| VGGNet-19 | CIFAR-100 | 90% | 160(S) | 74.61% | link |
| VGGNet-19 | CIFAR-100 | 95% | 30(S) | 71.76% | link |
| VGGNet-19 | CIFAR-100 | 95% | 160(S) | 73.35% | link |
| ResNet-32 | CIFAR-10 | 90% | 30(S) | 93.70% | link |
| ResNet-32 | CIFAR-10 | 90% | 160(S) | 94.39% | link |
| ResNet-32 | CIFAR-10 | 95% | 30(S) | 92.90% | link |
| ResNet-32 | CIFAR-10 | 95% | 160(S) | 93.41% | link |
| ResNet-32 | CIFAR-100 | 90% | 30(S) | 72.22% | link |
| ResNet-32 | CIFAR-100 | 90% | 160(S) | 73.43% | link |
| ResNet-32 | CIFAR-100 | 95% | 30(S) | 69.38% | link |
| ResNet-32 | CIFAR-100 | 95% | 160(S) | 70.31% | link |
| ResNet-50 | ImageNet | 80% | 30(S) | 75.19% | link |
| ResNet-50 | ImageNet | 90% | 30(S) | 72.43% | link |
To test our pruned models, download the pruned models and place them in the ckpt folder.
-
Select a configuration file in
configsto test the pruned models. For example, to evaluate a lottery jackpot for pruning ResNet-32 on CIFAR-10, run:python evaluate.py --config configs/resnet32_cifar10/evaluate.yaml --gpus 0To evaluate a lottery jackpot for pruning ResNet-50 on ImageNet, run:
python evaluate.py --config configs/resnet50_imagenet/evaluate.yaml --gpus 0