ComfyUI Node
Musubi Sampling Settings (Wan)
A ComfyUI node in musubi-tuner/wan/settings with 7 inputs and 1 output.
Musubi Sampling Settings (Wan)
- sampling_settings
◄sample_every_n_steps0►
◄sample_every_n_epochs0►
◄sample_at_firsttrue►
◄vae_name▾►
◄t5_name▾►
◄clip_nameNone►
◄prompts_texta beautiful sunset over mountains►
Categorymusubi-tuner/wan/settings
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| sample_every_n_steps | INT | 0 | Generate a sample every n steps. |
| sample_every_n_epochs | INT | 0 | Generate a sample every n epoch. |
| sample_at_first | BOOLEAN | true | Genera a sample at the starting training. |
| vae_name | COMBO | The same vae used to caching | |
| t5_name | COMBO | The same T5 used to caching | |
| clip_name | COMBO | None | If I2V, the same clip vision used for caching |
| prompts_text | STRING | a beautiful sunset over mountains | Write a prompt. You can specify width and height by using --w --h. Example : a cat making surf --w 832 --h 480 |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| sampling_settings | DICT | — |