search

mit-han-lab / efficientvit

EfficientViT is a new family of vision models for efficient high-resolution vision.
Apache License 2.0
1.79k stars 164 forks source link

SAM data #25

Closed daeyun closed 4 months ago

daeyun commented 1 year ago

Thanks for the great work. :raised_hands: How much data was the SAM model trained on? (e.g. all (or a subset) of SA-1B)