FromListGet1Latent¶
Documentation¶
- Class name:
FromListGet1Latent
- Category:
Bmad/Lists/Get1
- Output node:
False
This node is designed to extract a single latent representation from a list of latents. It simplifies the process of selecting a specific latent item for further operations or analysis, making it easier to work with collections of latent representations.
Input types¶
Required¶
list
- The list of latent representations from which a single latent is to be extracted. This parameter is crucial for specifying the source of the latent to be selected.
- Comfy dtype:
LATENT
- Python dtype:
List[torch.Tensor]
index
- The index of the latent representation to be extracted from the list. This parameter determines which item in the list is selected.
- Comfy dtype:
INT
- Python dtype:
int
Output types¶
latent
- Comfy dtype:
LATENT
- The single latent representation extracted from the provided list. This output is essential for downstream tasks that require a specific latent item.
- Python dtype:
torch.Tensor
- Comfy dtype:
Usage tips¶
- Infra type:
CPU
- Common nodes: unknown