The Beauty of
"Write once, train for anywhere"
😂
How long is that line?
Write Music
===
DATA
IN
DATA
OUT
DATA
IN
DATA
OUT
PLACES
Download Existing Models
Train a
Model
Modify an Existing
Model
Machine Learning & JavaScript is Awesome!
Decent but limited. It's a nice start, and with proper backing, it could become a Machine Learning standard library.
WEB
Free course on GitHub by Magnus Erik Hvass Pedersen
GitHub Photo
Your ALL-JS solution. Optimized and synced between GPU and CPU for indepth training. GPU is around the corner, and conv is too, in Brain 2.0
Robert Plummer
react-native-vision
5GB of images => 80MB model
REALTIME MODELS on your phone!?
This is epic! But the advantage is how you can run advanced models. Like generators.
How does one make a model?
You don’t need to come up with advanced algorithms anymore. You just have to teach a computer to come up with its own advanced algorithm.
Example Case
What's the hard part?
DATA
Nic or Nickelback
Feed Teacher
and let it train!
5+ Hours of This
Transfer Learning!
🎉
Think in ML
"Machine Learning does things a human can do;
trained with a lot of time and a lot of examples."
"Reimagining everything inspires science."
https://inspirobot.me/
#FML
Keep in touch:
https://infinite.red/
http://gantlaborde.com/
@GantLaborde Twitter