Neural network weights retain custom visual initializations after full training · Digg

/TECH

Neural network weights retain custom visual initializations after full training

The embedded watermarks did not degrade model training performance

Combined views

3.6K

2 posts, first seen 8h ago

16 likes

Neural network weights retain custom visual initializations after full training

The embedded watermarks did not degrade model training performance

Commentary on X

Highest ranked

arunabh@ArunabhMishra8TECH

Combined views

3.6K

2 posts, first seen 8h ago

16 likes1 comments

1 comments

1 saves

2 reposts

Sentiment

Positive0%100%Negative

Positive Read

Some users dismissed reports that neural net weights retain initialized face images after training via profanity and direct insults.

Based on 1 visible X reactions from 1 accounts; directional sample.

Digg Deeper

Ask a question below.

Published answers will appear here.

Commentary on X

Highest ranked

arunabh@ArunabhMishra8TECH

@vikhyatk @willdepue G-foff-nets

Gill Verdon@GillVerdTECH

The human urge to create things in our (literal) image https://twitter.com/willdepue/status/2076617090317078810

vik@vikhyatkTECH

@willdepue kaiming is out. will depue is in

@vikhyatk @willdepue G-foff-nets

Gill Verdon@GillVerdTECH

The human urge to create things in our (literal) image https://twitter.com/willdepue/status/2076617090317078810

vik@vikhyatkTECH

@willdepue kaiming is out. will depue is in

1 saves

2 reposts

Sentiment

Positive0%100%Negative

Positive Read

Some users dismissed reports that neural net weights retain initialized face images after training via profanity and direct insults.

Based on 1 visible X reactions from 1 accounts; directional sample.

Digg Deeper

Ask a question below.

Published answers will appear here.