¿Tienes un flujo de trabajo de ComfyUI? Obtén un Workflow Doc legible en tu navegador.Haz legible un flujo de trabajo.Pruébalo ahora ↗
Simple WAN T2V Workflow for Self Forcing
Perfil de ejecución
Descripción de la fuente
Simple WAN T2V Workflow for Self Forcing
Self Forcing trains autoregressive video diffusion models by simulating the inference process during training, performing autoregressive rollout with KV caching. It resolves the train-test distribution mismatch and enables real-time, streaming video generation on a single RTX 4090 while matching the quality of state-of-the-art diffusion models.
Update (i2v):
To use Vace, you will need to use a different checkpoint: https://huggingface.co/lym00/Wan2.1-T2V-1.3B-Self-Forcing-VACE/blob/main/Wan2.1-T2V-1.3B-Self-Forcing-DMD-VACE-FP16.safetensors
Comentario generado por IA
Explicación generada por IA basada en los detalles de la fuente y la configuración. Las sugerencias están claramente identificadas.
This workflow is for generating video with Self Forcing and the Wan Video 1.3B t2v base model. It supports T2V, I2V, and V2V workflows, using text, images, or video as inputs and producing video.
Self Forcing is described as autoregressive rollout with KV caching that simulates inference during training.
The documented task forms are T2V, I2V, and V2V, with video as the output.
Find it on Civitai at https://civitai.com/models/1668005; the listed publisher is Powerpuff.
It requires the `comfyui-videohelpersuite` node pack.
The listed manual model files are `umt5_xxl_fp8_e4m3fn_scaled.safetensors`, `wan_2.1_vae.safetensors`, and `Wan2.1-T2V-1.3B-Self-Forcing-DMD-VACE-FP16.safetensors`.
For Vace, use `Wan2.1-T2V-1.3B-Self-Forcing-DMD-VACE-FP16.safetensors` as the different checkpoint, and use `self_forcing_dmd.pt` as the T2V checkpoint.
Verify that the three listed model files are available in your setup before using the workflow.
Sugerencia · sin verificar
If you plan to use Vace, verify that the Vace checkpoint and T2V checkpoint match the task you want to run.
Sugerencia · sin verificar
¿Es necesario editarlo?
Inicia sesión para solicitar una edición.
Fuentes
1 fuenteExtractos de fuentes
1 extractoSource context: 5945 downloads · Type Workflows · Base model Wan Video 1.3B t2v
Download self_forcing_dmd.pt from https://huggingface.co/gdhe17/Self-Forcing/tree/main/checkpoints and use it as the t2v checkpoint.
Project website: https://self-forcing.github.io/
Requisito de VRAM estimado
Estimación no disponible
Los tamaños de los modelos aún no están lo bastante completos para calcular este requisito.
Requisitos
4 requisitosumt5_xxl_fp8_e4m3fn_scaled.safetensors
Archivo sin verificarText encoder · SAFETENSORS · Unknown
VAE · Unknown
Wan2.1-T2V-1.3B-Self-Forcing-DMD-VACE-FP16.safetensors
No resueltoCheckpoint · Unknown
Paquete de nodos · Registry
