tobran / GALIP

[CVPR2023] A faster, smaller, and better text-to-image model for large-scale training
MIT License
225 stars 25 forks source link

Adaption for different variant of CLIP #7

Open CycAreal opened 1 year ago

CycAreal commented 1 year ago

Hi MingTao, thanks for the wonderful work! I am trying to adapt the GALIP for larger variants of CLIP such as CLIP-L/14, but got trouble modifying the NetC and NetD, have you experimented with this? Could you please give me some advice on the modification?