Downloads · 30 days
0
yjitiu/ComfyUI-Binyuan-Latent-Helper
ComfyUI-Binyuan-Latent-Helper is a machine learning model from yjitiu. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
A tiny ComfyUI custom node that generates an empty LATENT (and outputs the matching width/height) from a list of Chinese-labeled size presets — so you don't have to remember "1216×832" etc. Bilingual UI.
Downloads · 30 days
0
Access
Public
Updated Jul 1, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
.md3.4 KB · 34%
From the Hugging Face model README
A tiny ComfyUI custom node that generates an empty LATENT (and outputs the matching width/height) from a list of Chinese-labeled size presets — so you don't have to remember "1216×832" etc. Bilingual UI.
这是一个极简的单节点插件:用一个下拉预设快速生成空 Latent,同时把实际宽高作为整数端口输出,方便下游节点(如采样器)直接读取尺寸。带前端 JS,预设与宽高输入框联动。
节点 📏 binyuan · 中文Latent快捷助手 v2.0,分类 Latent助手。
按预设或自定义尺寸生成一个全零的空 Latent 张量(4 通道,[批次, 4, 高/8, 宽/8]),并把实际宽高作为 INT 端口输出。
ComfyUI/custom_nodes/binyuan_latent_helper/,重启 ComfyUI。Latent助手 → 📏 binyuan · 中文Latent快捷助手 v2.0。1024x1024 (1:1 正方形))或选 自定义 后手填宽高。Latent 端口接到采样器的 latent 输入;宽度/高度 可接到需要尺寸的节点。| 参数 | 说明 |
|---|---|
| 预设选择 | 一组常用尺寸(正方形/竖屏/横屏/手机壁纸/2K/老模型 512 等)+ 自定义 |
| 宽度 / 高度 | 64–8192,步长 8 |
| 批次大小 | 1–64 |
| 宽高对调 | 开启后宽高互换,输出的 宽度/高度 端口也同步为对调后的值 |
Latent:空 Latent(LATENT)宽度 / 高度:实际生效的宽高(INT,对调时为互换后的值)生成的是 4 通道空 Latent,适合 SD1.5 / SDXL 等 4 通道模型;Flux 等 16 通道模型请用对应加载器自带的空 Latent。

Node 📏 binyuan · 中文Latent快捷助手 v2.0 under category Latent助手.
Generates an all-zero empty Latent ([batch, 4, H/8, W/8]) from a preset or
custom size, and also returns the effective width/height as integer ports.
ComfyUI/custom_nodes/binyuan_latent_helper/ and restart ComfyUI.Latent助手 → add the node.1024x1024 (1:1 正方形)) or choose 自定义 and type W/H.Latent port to a sampler's latent input; 宽度/高度 feed
any node that needs the size.预设选择 (preset list), 宽度/高度 (64–8192, step 8), 批次大小 (1–64),
宽高对调 (swap W/H — outputs reflect the swapped values).
Latent (LATENT), 宽度/高度 (INT — the effective size after any swap).
Produces a 4-channel empty Latent (SD1.5 / SDXL style). For 16-channel models like Flux, use the loader's own empty-Latent node.
__init__.py — node definition & registration (WEB_DIRECTORY = "js")js/chicun.js — frontend: syncs the preset dropdown with W/H inputsREADME.md / LICENSE (MIT)Binyuan Latent Helper,或手动 git clone 到 custom_nodes/。git clone https://github.com/yjitiu/ComfyUI-Binyuan-Latent-Helper.git binyuan_latent_helper