Learning-Rate-Dropout
Pytorch implementation of Learning Rate Dropout.
Paper Link: https://arxiv.org/pdf/1912.00144.pdf
Train ResNet-34 for Cifar10:
run:
python main.py --model=resnet --optim=adam_lrd --lr=0.001 --LRD_p=0.5
python main.py --model=resnet --optim=adam --lr=0.001
python main.py --model=resnet --optim=sgd_lrd --lr=0.1 --LRD_p=0.5
python main.py --model=resnet --optim=sgd --lr=0.1
python main.py --model=resnet --optim=rmsprop_lrd --lr=0.001 --LRD_p=0.5
python main.py --model=resnet --optim=rmsprop --lr=0.001
After training, run "plot.py" to show the learning curves.
701 Dec 30, 2022
62 Dec 08, 2022
610 Dec 28, 2022
114 Dec 20, 2022
46 Nov 10, 2022
1.2k Dec 27, 2022
56 Nov 28, 2022
21 Nov 22, 2022
104 Jan 04, 2023
1.9k Jan 06, 2023
345 Dec 28, 2022
727 Jan 01, 2023
2.1k Jan 02, 2023
3 Nov 09, 2021
2 Apr 07, 2022
2.2k Jan 08, 2023
9 Oct 29, 2022
643 Dec 30, 2022
569 Dec 30, 2022
10 Jun 23, 2022