-
Notifications
You must be signed in to change notification settings - Fork 306
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Share some fine-tuned hyperparameter settings #136
Comments
fedavg: |
per-fedavg python main.py -nb 100 -data Cifar100 -m cnn -algo PerAvg -gr 1000 -did 0 -lr 0.005 -bt 0.001 -go cnn > cifar100_per-fedavg_1.out 2>&1 python main.py -nb 200 -data Tiny-imagenet -m cnn -algo PerAvg -gr 1000 -did 0 -lr 0.005 -bt 0.001 -go cnn > tiny_per-fedavg_1.out 2>&1 |
APFL: python main.py -nb 100 -data Cifar100 -m cnn -algo APFL -gr 1000 -did 0 -lr 0.005 -al 1.0 -go cnn > cifar100_APFL_1.out 2>&1 python main.py -nb 200 -data Tiny-imagenet -m cnn -algo APFL -gr 1000 -did 0 -lr 0.005 -al 1.0 -go cnn > tiny_APFL_1.out 2>&1 |
pFedMe python main.py -nb 100 -data Cifar100 -m cnn -algo pFedMe -gr 1000 -did 0 -lr 0.005 -lrp 0.01 -bt 1 -lam 15 -K 5 -go cnn > cifar100_pFedMe_1.out 2>&1 python main.py -nb 200 -data Tiny-imagenet -m cnn -algo pFedMe -gr 1000 -did 0 -lr 0.005 -lrp 0.01 -bt 1 -lam 15 -K 5 -go cnn > tiny_pFedMe_1.out 2>&1 |
-gr不是2000吗? |
1000轮内就完全可以收敛了 |
是的, |
FedAvg真的可以在1000个round下收敛吗?我在“serveravg.py”中加了这两句: 顺便一提,按照上面提到的pFedMe的超参数设定(和ALA文中提到的一致),似乎完全没有办法收敛,即便跑完2000rounds,下面是我pFedMe的log 结果 超参数及模型 ============= Running time: 0th ============= Join ratio / total clients: 1.0 / 20 |
为什么你们的cnn是小写的?main.py代码上是 |
@KalosAner 代码会持续迭代优化的。所以之前的超参数也仅供参考,建议重新调参 |
原来如此,好的 |
以下是我使用该仓库进行的一些方法复现所使用的超参数,走过路过的朋友们,如果发现哪里超参数的选择不合理或该使用的超参数未使用到,麻烦提醒下我,不胜感激。
The text was updated successfully, but these errors were encountered: