- 版本 [version]
- 下载 112
- 文件大小 13052
- 文件计数 1
- 创建日期 2026年7月29日
- 最后更新 2026年9月21日
02.文生图高清放大(Upscale Image)
02.文生图高清放大(Upscale Image):基于 SD 1.5 搭建的高清放大方案,节点规模 12 个、连线 16 条。工作流主路径为 CLIPTextEncode ×2、UpscaleModelLoader ×1、CheckpointLoaderSimple ×1,配合 输出图像,默认输出 512×768;节点已按执行顺序排布,便于二次修改。主模型 4x-UltraSharp.pth、SD15majicmixRealistic_v7.safetensors、SD15vaeFtMse840000Ema_v10.safetensors。下载后直接导入即可使用,建议显存 4GB 起。
- 主模型:4x-UltraSharp.pth、SD15majicmixRealistic_v7.safetensors、SD15vaeFtMse840000Ema_v10.safetensors
- 核心节点:CLIPTextEncode ×2、UpscaleModelLoader ×1、CheckpointLoaderSimple ×1
- 输出形式:输出图像 · 512×768
- 规模:12 节点 / 16 连线
- 建议显存:4GB 起
- 适用场景:2K/4K/8K 无损放大、细节重绘与纹理修复,救回模糊与 AI 感素材。
工作流节点图
图较大可横向滚动;点击图片查看高清矢量原图(浏览器中可自由缩放)。
资源头图

使用说明
下载工作流文件后在 ComfyUI 中直接拖入画布(或菜单 Load 导入)即可运行;缺失的自定义节点用 ComfyUI Manager 一键安装,模型按文件名放入对应的 models 目录。
源码预览
{
"last_node_id": 36,
"last_link_id": 73,
"nodes": [
{
"id": 26,
"type": "UpscaleModelLoader",
"pos": [
1132.701171875,
667.3887939453125
],
"size": [
315,
58
],
"flags": [],
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "UPSCALE_MODEL",
"type": "UPSCALE_MODEL",
"links": [
46
],
"shape": 3,
"label": "UPSCALE_MODEL"
}
],
"properties": {
"Node name for S&R": "UpscaleModelLoader"
},
"widgets_values": [
"4x-UltraSharp.pth"
]
},
{
"id": 2,
"type": "CheckpointLoaderSimple",
"pos": [
-682.311279296875,
519.2742919921875
],
"size": [
315,
98
],
"flags": [],
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
1,
42
],
"shape": 3,
"label": "MODEL"
},
{
"name": "CLIP",
"type": "CLIP",
"links": [
8,
9
],
"shape": 3,
"label": "CLIP"
},
{
"name": "VAE",
"type": "VAE",
"links": [],
"shape": 3,
"label": "VAE"
}
],
"properties": {
"Node name for S&R": "CheckpointLoaderSimple"
},
"widgets_values": [
"SD15majicmixRealistic_v7.safetensors"
]
},
{
"id": 1,
"type": "KSampler",
"pos": [
304.59368896484375,
493.2742919921875
],
"size": [
315,
262
],
"flags": [],
"order": 7,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 1,
"slot_index": 0,
"label": "model"
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 2,
"slot_index": 1,
"label": "positive"
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 3,
"slot_index": 2,
"label": "negative"
},
{
"name": "latent_image",
"type": "LATENT",
"link": 4,
"slot_index": 3,
"label": "latent_image"
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
5
],
… (完整内容请下载文件查看)
代码较长,以上仅前 3800 字符 —— 点击下方按钮下载完整 JSON。



