-
Notifications
You must be signed in to change notification settings - Fork 1.2k
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
[飞桨黑客松第四期] (128) MobileViTv3 #2691
Merged
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Thanks for your contribution! |
This reverts commit 35fc732.
modified batch_size per gpu and update_freq in MobileViTv3_S.yaml for training with 4 gpus
TingquanGao
approved these changes
Apr 6, 2023
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
TODO: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
任务详情: [飞桨黑客松第四期任务总览] Paddle#51281
修改内容
[bugfix]当DecodeImage
使用pil
backend时,image类型应转为RGB[update] 在ExponentialMovingAverage
中使用numpy来完成ema weight的更新,可以一定程度上提速(大概快一倍)MixupOperator
、CutmixOperator
和FmixOperator
中,索引打乱的方法存在一定缺陷,np.random.permutation
可能造成部分索引在打乱前后顺序不变。改为逆序打乱索引。[update] 为RandCropImage
添加use_log_aspect
的支持MultiScaleSampler
中epoch
不自增的bugepoch
为当前的epochRandAugmentV3
,用于MobileViTv3-v2模型的训练复现精度
测试
top1
top5
(G)
(M)
训练
训练平台AIStudio:python3.7 + paddle2.4.0 + cuda11.2 + 4 * V100 32GB
top1
top5
权重及训练日志下载地址:百度网盘