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.
8 Dec 27, 2022
20 Jul 18, 2022
5 Feb 04, 2022
16 Oct 16, 2022
55 Nov 14, 2022
6.2k Jan 09, 2023
36 Dec 21, 2022
6 May 03, 2022
5 Oct 31, 2022
579 Dec 22, 2022
46 Dec 15, 2022
27 Jan 07, 2023
22 Aug 23, 2022
405 Jan 02, 2023
7 Nov 22, 2022
8 Jun 23, 2021
21 Nov 05, 2022
23 Dec 13, 2022
51 Nov 20, 2022
281 Dec 21, 2022