nunchaku
nunchaku copied to clipboard
load the int4 schnell model spend 190s for 1.04
trafficstars
190S
How much VRAM do you have? You can try disabling the cpu_offload for better speed.
See our FAQ:
Try upgrading your CUDA driver and try setting the environment variable
NUNCHAKU_LOAD_METHODto eitherREADorREADNOPIN. Some references: #249 #311 #276
I'll close this issue for now. If the error persists or you encounter any other problems, feel free to open a new issue.