A Diffusers snippet?
#1
by tintwotin - opened
Could you share a Diffusers snippet indicating how to use these weights? Do you load it as a custom pipeline?
Normally, to load SDNQ prequantized model you'd simply install and register sdnq and then load as usual.
pip install sdnq
from sdnq import SDNQConfig # import sdnq to register it into diffusers and transformers
pipe = diffusers.xxxxPipeline.from_pretrained(repo_id)
for Anima, SD.Next has a custom pipeline as diffusers does not implement it fully.
Pipeline code is included here in pipeline.py