Diffusion Models (choose one based on your hardware capabilities):
Lightweight option: SkyReels-V2-I2V-1.3B-540P
for image to video Download:
for text to video
https://civarchive.com/models/1497893?modelVersionId=1694508
Place in:
ComfyUI/models/diffusion_models/
CLIP Vision Model:this for image to video
clip_vision_h.safetensors
Place in:
ComfyUI/models/clip_vision/
Text Encoder Models:
Download from: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/tree/main/split_files/text_encoders
Place in:
ComfyUI/models/text_encoders/
VAE Model:
https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/tree/main/split_files/vae
wan_2.1_vae.safetensors
Download:
Place in: ComfyUI/models/vae/
Description
FAQ
Comments (10)
;)
this model (skyreels) is compatible with wan loras ? o is different?
i tried one but has no effect
plz try A feed back
Yeah, it does work. In some of my tests, had luck using wan2.1 loras with the same base model size (14B loras with a 14B model). But when 14B loras are used with a 1.3B model, there’s no noticeable effect.
@lym0 wow that intrested , so lora must be trained on 1.3 b model to be worked
i was testing with some loras of 14b with strength > 1.5 and this take effect, but not all are compatibles
is this usable without a llm? And can i Use GGUF versions?
u can yes use gguf insted of normal one also u can rm prompt generation and write urs manually
You can remove all llm prompt nodes, still works