Inference settings for VBVR-wan2.2-comfy-bf16.safetensors
1
#3 opened about 1 month ago
by 
Heouzen
is the “SNR-Calibrated-Hybrid”I in the document an HiFi FP8 model, and does it need to be used with lightx2Vlora for acceleration?
2
#2 opened about 2 months ago
by 
RedHn
Can this model be quantized to FP8 scale?
8
#1 opened 3 months ago
by 
Nuke1229