Resources you need:
đź“‚Files :
For base version
I2V Model : wan2.2_i2v_high_noise_14B_fp8_scaled.safetensors and wan2.2_i2v_low_noise_14B_fp8_scaled.safetensors
In models/diffusion_models
CLIP: umt5_xxl_fp8_e4m3fn_scaled.safetensors
in models/clip
For GGUF version
I2V Quant Model : wan2.2_i2v_high_noise_14B_QX.gguf and wan2.2_i2v_low_noise_14B_QX.gguf
In models/unet
Quant CLIP: umt5-xxl-encoder-QX.gguf
in models/clip
For lightx2V version
lightx2V LoRA : Wan21_I2V_14B_lightx2v_cfg_step_distill_lora_rank64.safetensors
VAE: wan_2.1_vae.safetensors
in models/vae
ANY upscale model:
Realistic : RealESRGAN_x4plus.pth
Anime : RealESRGAN_x4plus_anime_6B.pth
in models/upscale_models
📦Custom Nodes :
Description
Base version
FAQ
Comments (8)
🤩
Got an error saying it "cannot access local variable 'clip_vision_output' where it is not associated with a value".
Maybe I'm missing something...?
Strange i dont use any clip vision for WAN2.2, you have updated comfyui?
UmeAiRTÂ You're right! The problem was on my side. I was on Comfy's v0.3.47 instead of v0.3.48. Everything appears to be working now! Thank you for sharing the workflow!
For anyone else reading: I also had to update the custom node Lora Manager. They released an update yesterday and without it you'll be getting some weird errors.
Laura lightx2V -LoRA (i2v) does not work normally, you need to install text t2v
10/10 workflow
This workflow is amazing like always! :)
Gotta add the MagCache custom nodes to the requirements in the description tho 🙏
@UmeAiRT
Is there a way to set the weight/balance of the 1st or last frames? so that it doesn't just blending into the last frame in a few frames?