We introduce a novel approach for parameter generation, named neural network parameter diffusion (p-diff), which employs a standard latent diffusion model to synthesize a new set of parameters
I can see another (more recent) paper has been linked to this repo as well, Conditional LoRA Param Generation. Do you guys have any instructions/code released in this repo? The work looks exciting and it would be great to play with 👀
Hi guys,
I can see another (more recent) paper has been linked to this repo as well, Conditional LoRA Param Generation. Do you guys have any instructions/code released in this repo? The work looks exciting and it would be great to play with 👀