• 版本 [version]
• 下载 78
• 文件大小 28166
• 文件计数 1
• 创建日期 2026年4月28日
• 最后更新 2026年9月21日

Nano-Banana+wan2_2_14B_I2V _4步加速

这是一份开箱即用的视频生成工作流（Wan 2.2、Wan 2.1、Nano Banana），拖入 ComfyUI 即可运行，全图 28 节点 / 36 连线。核心流程 ModelPatchTorchSettings ×2、PathchSageAttentionKJ ×2、CLIPTextEncode ×2，输出视频，参数已调好，替换输入图即可复现效果。主模型 Wan2.2Wan2.2-I2V-A14B-HighNoise-Q8_0.gguf、Wan2.2lightx2v_i2v_low_noise_model.safetensors、Wan2.2lightx2v_i2v_high_noise_model.safetensors。已挂载 LoRA：Wan2.2lightx2v_i2v_low_noise_model.safetensors、Wan2.2lightx2v_i2v_high_noise_model.safetensors。下载后直接导入即可使用，建议显存 16GB 起（视频生成建议 24GB）。

• 主模型：Wan2.2Wan2.2-I2V-A14B-HighNoise-Q8_0.gguf、Wan2.2lightx2v_i2v_low_noise_model.safetensors、Wan2.2lightx2v_i2v_high_noise_model.safetensors
• LoRA / 适配器：Wan2.2lightx2v_i2v_low_noise_model.safetensors、Wan2.2lightx2v_i2v_high_noise_model.safetensors
• 核心节点：ModelPatchTorchSettings ×2、PathchSageAttentionKJ ×2、CLIPTextEncode ×2
• 输出形式：输出视频
• 规模：28 节点 / 36 连线
• 建议显存：16GB 起（视频生成建议 24GB）
• 适用场景：图生视频、首尾帧、数字人与长视频一致性控制工作流。

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

{
    "id": "ec7da562-7e21-4dac-a0d2-f4441e1efd3b",
    "revision": 0,
    "last_node_id": 89,
    "last_link_id": 182,
    "nodes": [
        {
            "id": 72,
            "type": "ModelPatchTorchSettings",
            "pos": [
                629.07421875,
                -263.69696044921875
            ],
            "size": [
                307.443359375,
                58
            ],
            "flags": [],
            "order": 20,
            "mode": 0,
            "inputs": [
                {
                    "label": "model",
                    "localized_name": "model",
                    "name": "model",
                    "type": "MODEL",
                    "link": 153
                },
                {
                    "label": "enable_fp16_accumulation",
                    "localized_name": "enable_fp16_accumulation",
                    "name": "enable_fp16_accumulation",
                    "type": "BOOLEAN",
                    "widget": {
                        "name": "enable_fp16_accumulation"
                    },
                    "link": null
                }
            ],
            "outputs": [
                {
                    "label": "MODEL",
                    "localized_name": "模型",
                    "name": "MODEL",
                    "type": "MODEL",
                    "links": [
                        156
                    ]
                }
            ],
            "properties": {
                "cnr_id": "comfyui-kjnodes",
                "ver": "40d9d84b68939b09d61d6a41ac15ceef8c0b8d22",
                "Node name for S&R": "ModelPatchTorchSettings"
            },
            "widgets_values": [
                true
            ]
        },
        {
            "id": 8,
            "type": "VAEDecode",
            "pos": [
                2165.319580078125,
                -424.4515380859375
            ],
            "size": [
                210,
                46
            ],
            "flags": [],
            "order": 26,
            "mode": 0,
            "inputs": [
                {
                    "label": "samples",
                    "localized_name": "Latent",
                    "name": "samples",
                    "type": "LATENT",
                    "link": 124
                },
                {
                    "label": "vae",
                    "localized_name": "vae",
                    "name": "vae",
                    "type": "VAE",
                    "link": 76
                }
            ],
            "outputs": [
                {
                    "label": "IMAGE",
                    "localized_name": "图像",
                    "name": "IMAGE",
                    "type": "IMAGE",
                    "slot_index": 0,
                    "links": [
                        157
                    ]
                }
            ],
            "properties": {
                "cnr_id": "comfy-core",
                "ver": "0.3.45",
                "Node name for S&R": "VAEDecode"
            },
            "widgets_values": []
        },
        {
            "id": 76,
            "type": "VHS_VideoCombine",
            "pos": [
                2424.410400390625,
                -424.4515380859375
            ],
            "size": [
                220.5830078125,
                701.570068359375
            ],
            "flags": [],
            "order": 27,
            "mode": 0,
            "inputs": [
                {
                    "label": "images",
                    "localized_name": "images",
                    "name": "images",
                    "type": "IMAGE",
                    "link": 157
                },
                {
                    "label": "audio",
             
… (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。