Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
How to push custom model
#60
by
mirth
- opened
I need clarification about how HF works.
Suppose I've made custom text embedding model with transformer as backbone but also with some custom layers.
What happen If i push such model to HF? How it handle my custom layers?
You can push pretty much anything. It may not be loadable with transformers but you can provide your own loading instructions; For adding a model to MTEB you do not even need to push it to HF if you don't want to