What are you trying to do?
Create images using z.image turbo with this faster model and his kernels (nunchaku int4)
What have you tried?
Install and reinstall this models, but seems not work, the dev console say
"Loading Model 'ckpts\svdq-int4_r256-z-image-turbo.safetensors' ...
Loading Text Encoder 'ckpts\Qwen3\qwen3_quanto_bf16_int8.safetensors' ...
[nunchaku_int4] Nunchaku kernels unavailable; using Python fallback.
Pinning data of 'transformer' to reserved RAM"
As i read this kernels are very fast and optimized for hadware like mine and rtx 3000 series, but speed still the same as other z turbo models i tried :( . BTW this WAN2GP is awesome, i really love it <3
Your system (OS / GPU / RAM / VRAM / etc.)
Windows 10 / RTX 3050 / 24 GB RAM / 8 GB VRAM / SSD KINGSTON 400GB/
What are you trying to do?
Create images using z.image turbo with this faster model and his kernels (nunchaku int4)
What have you tried?
Install and reinstall this models, but seems not work, the dev console say
"Loading Model 'ckpts\svdq-int4_r256-z-image-turbo.safetensors' ...
Loading Text Encoder 'ckpts\Qwen3\qwen3_quanto_bf16_int8.safetensors' ...
[nunchaku_int4] Nunchaku kernels unavailable; using Python fallback.
Pinning data of 'transformer' to reserved RAM"
As i read this kernels are very fast and optimized for hadware like mine and rtx 3000 series, but speed still the same as other z turbo models i tried :( . BTW this WAN2GP is awesome, i really love it <3
Your system (OS / GPU / RAM / VRAM / etc.)
Windows 10 / RTX 3050 / 24 GB RAM / 8 GB VRAM / SSD KINGSTON 400GB/