Filip & DeFILIP? :)
christophschuhmann opened this issue · comments
Hi, I just realized that you have Filip and DeFilip implementations, very interesting.
Do you already have results How they compare with Clip and DeCLIP?
With respect benchmark scores and compute efficiency? :)
May I also ask about the details of the hardware you used for training?
… And - do you have code for multinode training?
We will publish a technical report on arxiv within two weeks and the result and checkpoints will be released in the same time.
The code has already supported multi-node training, but the ddp framework is being debugged.
Any update as to when FILIP weights may be released?
The FILIP and DeFILIP weights are released new.