Nvidia Launches a 100kb text-to-image model called Perfusion
research.nvidia.com
Nvidia Launches a 100kb text-to-image model called Perfusion
1–10 of 19 posts
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#2* a several GB model to which a small amount of subject specific training can be performed on archetype, augmenting that large model
Doesn’t seem to be any code or runtime examples
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#3Very misleading title
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#4[flagged]
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#5It's not a 100kb model. It's 100kb config files for a several GB model. A small trained layer to stick on top of the real model for fine tuning.
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#6Very misleading title
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#7[flagged]
They can’t, and they aren’t. In any case as you’ve just read in the article the model is very large, why even say this given what you just read?
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#8It's not a 100kb model. It's 100kb config files for a several GB model. A small trained layer to stick on top of the real model for fine tuning.
Yes, although it is decently interesting that a model can be fine tuned by just tweaking a small number of weights and training for just a few minutes
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#9Maybe the title should be the "Key-Locked Rank One Editing for
Text-to-Image Personalization", per HN guidelines.
Re: Nvidia Launches a 100kb text-to-image model called Perfusion
#10Yes, you need the pretrained model. BUT: for embedded applications, you could put that into metal and have the 100kb in flash which could open up some possibilities.