Maxout uses the max-norm constraint (or dropout does). Do you know when to enforce the constraint, during update ? or during calculating the pre-activation?
[link][3 comments]
submitted by rishok
[link][3 comments]
↧
Channel:
Machine Learning
X
Are you the publisher?
Claim or
contact us
about this channel.
X
0
Channel Details:
- Title: Machine Learning
- Channel Number: 1745337
- Language: English
- Registered On: November 24, 2011, 3:47 am
- Number of Articles: 63175
- Latest Snapshot: June 6, 2025, 1:49 am
- RSS URL: http://www.reddit.com/r/machinelearning/.rss
- Publisher: https://www.reddit.com/r/machinelearning/
- Description:
- Catalog: //unbudded.rssing.com/catalog.php?indx=1745337
© 2025 //www.rssing.com