Y
Hacker News
new
|
ask
|
show
|
jobs
by
arivero
1200 days ago
the idea is to do a mininal training on an existing model, so minimal addition of new tokens