Re:Exploiting machine learning Pickle files
by mrguyorama on Thursday March 18, 2021 @01:55PM
(#26505480)
Our team is literally running into this right now. XGBoost claims their non-pickle implementation of model file (which is just json) is "experimental"
[ Parent
| on HN ]
Re:Exploiting machine learning Pickle files
by a-dub on Thursday March 18, 2021 @01:58PM
(#26505514)
been there, done that (json parameter storage). it's slow as molasses and precision loss from storing floats as text can cause issues.
Re:Exploiting machine learning Pickle files
by craigacp on Thursday March 18, 2021 @09:29PM
(#26509391)
XGBoost also has a binary file format which is neither pickle nor json. It's shared across their Python, Java and R frontends and parsed by the C++ library.
Re:Exploiting machine learning Pickle files
by mrguyorama on Friday March 19, 2021 @01:40PM
(#26516158)
Except the library claims that will be moved away from. To the JSON system