Based on hunyuan video wrapper by kijai.
Runtime profile
Source description
Based on hunyuan video wrapper by kijai.
Update 1/1/25
This adds device selection nodes to hunyuan video wrapper.
I have included the workflow.
AND there is a download for the changed files if you want to just manually replace the 2 files in /custom_nodes/Comfyui-HunyuanVideoWrapper/
Or you can download the git from here.
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Hunyuan Video MultiGPU Setup is a Hunyuan Video workflow for generating video on multi-GPU systems. It uses text encoding and the listed checkpoint, VAE, and LoRA inputs, then samples, decodes, and combines video; device selection is optional.
Its named stages cover model, VAE, and text encoder loading; text encoding; LoRA selection and editing; sampling; decoding; and video combining.
The workflow lists these model files: hunyuan_video_FastVideo_720_fp8_e4m3fn.safetensors, hunyuan_video_vae_bf16.safetensors, and HYNV-Ellie-Lora01.safetensors.
HyVideoDecode and VHS_VideoCombine are its named video decoding and combining components.
To add the wrapper changes, replace two files in /custom_nodes/Comfyui-HunyuanVideoWrapper/ or download the ComfyUI-HunyuanVWMultiGPU repository.
Named node components include DownloadAndLoadHyVideoTextEncoder, HyVideoCudaSelect, HyVideoDecode, HyVideoLoraBlockEdit, HyVideoLoraSelect, and HyVideoModelLoader.
Other named node components are HyVideoSampler, HyVideoTextEncode, HyVideoVAELoader, and VHS_VideoCombine.
On a multi-GPU system, use HyVideoCudaSelect if you want to choose a device; device selection is optional.
Suggestion · not verified
Before running, check that the three listed model files and named node components are available in your setup.
Suggestion · not verified
No license is listed; check the usage terms before reuse.
Suggestion · not verified
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
1 excerptSource context: 641 downloads · Type Workflows · Base model Hunyuan Video
This has be recoded from the last version to make this all optional, so I could get the pull request approved.
so this IS coming to base hunyuanvideowrapper.
I've modified the nodes to allow selection of device for those with multiple gpu's, this was necessary for me as my system has 2 4090's on a 1x pcie switch, so unloading and loading was slower then generating, with 2 4090's and the 4080 I don't have to unload anything during generation anymore.
Estimated VRAM requirement
Estimate unavailable
Model file sizes are not complete enough to calculate this requirement yet.
Requirements
13 requirementsDownloadAndLoadHyVideoTextEncoder
File unverifiedText encoder · Unknown · ComfyUI-HunyuanVideoWrapper
hunyuan_video_FastVideo_720_fp8_e4m3fn.safetensors
Not resolvedCheckpoint · Unknown
hunyuan_video_vae_bf16.safetensors
PossibleVAE · Unknown
HYNV-Ellie-Lora01.safetensors
Not resolvedLoRA · Unknown
VAE · Unknown · ComfyUI-HunyuanVideoWrapper
HyVideoCudaSelect
Not resolvedNode pack · Unknown
Node pack · Unknown
Node pack · Unknown
Node pack · Unknown
Node pack · Unknown
Node pack · Unknown
Node pack · Unknown
VHS_VideoCombine
PossibleNode pack · Unknown