未验证 提交 5d76f932 编写于 作者: J Jason 提交者: GitHub

Merge pull request #301 from FlyingQianMM/develop_qh

reset lr for ppyolo tutorial
...@@ -52,7 +52,7 @@ model.train( ...@@ -52,7 +52,7 @@ model.train(
train_dataset=train_dataset, train_dataset=train_dataset,
train_batch_size=8, train_batch_size=8,
eval_dataset=eval_dataset, eval_dataset=eval_dataset,
learning_rate=0.0005, learning_rate=0.000125,
lr_decay_epochs=[210, 240], lr_decay_epochs=[210, 240],
save_dir='output/ppyolo', save_dir='output/ppyolo',
use_vdl=True) use_vdl=True)
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册