logo
happy_move/cnb/comfyui
Public
0
0
Login
CodeIssuesPull requestsEventsPackagesInsights
main
Branch
1
Tag
0
Forkfromblueskyfff/comfyui6, ahead:main2 commits, behind:main1 commits
旭
Update
1d2e8ecd
4 commits
.ide
comfyui
custom_nodes
input
models
output
user
venv
.cnb.yml
.gitattributes
.gitignore
README.md
dockerfile
初始化下载

云原生构建的ComfyUI运行环境

基于nvidia/cuda镜像，以build模式生成开发环境（预装python3.12、vscode中文语言包、aria2c下载工具等）。之后生成venv并预装torch、flash_attn、sageattention等库。预装部分ComfyUI热门插件，并提供基于Gradio的图形化下载工具。

目前配置为从镜像启动。如需要每次启动重新构建docker环境可以自行修改 .cnb.yml 文件，重新构建的环境不含venv和ComfyUI，可参考venv生成方法和安装ComfyUI及插件。

相关项目链接：ComfyUI 在线体验（飞桨AI Studio星河社区）

最新功能

新增模型自动下载功能，ComfyUI启动后界面内可选择的模型在运行时会自动下载到云节点缓存，无需人为操作，也不需要保存到个人仓库！！！

详细代码实现见ComfyUI源码修改，开源项目欢迎复制/参考代码，但希望注明代码出处~

注意：

  1. 自动下载功能是修改ComfyUI源码实现的，如果自行更新ComfyUI，可能会失去自动下载功能！！！

  2. 目前自动下载功能包括checkpoints、clip、clip_vision、controlnet、diffusion_models、loras、text_encoders、upscale_models、vae文件夹~

  3. 对于插件需要的模型暂不支持自动下载，主要是插件有自己的模型查找方式，不一定调用ComfyUI底层提供的模型查找函数。

  4. 预置模型来源于AI Models、ComfyUI-yi_dian_tong，不可商用！不可商用！不可商用！

运行指南

1. 启动ComfyUI

环境启动后会自动启动ComfyUI。如需关闭，终端中按Ctrl+c。

再次启动ComfyUI，终端输入 qd，回车运行。可添加其他启动参数，如 qd --use-sage-attention，启用sageattention。

注意：--listen、--auto-launch参数已内置，无需额外添加。

2. 打开ComfyUI界面

四种打开外链的方法（选一种即可）：

  1. 环境启动后会自动解压生成venv并启动ComfyUI，会弹出一个对话框，点打开；
  2. 点击右下角弹出的对话框中的在浏览器中打开；
  3. 按住Ctrl键点击终端输出内容中的http://0.0.0.0:8188；
  4. 在端口下，按住Ctrl键点击外部链接。

3. 打开Gradio下载工具界面

首先点击右下角 + 打开第二个终端，在新终端中输入 ui，回车运行，将自动打开下载工具界面。如需关闭，对应终端中按Ctrl+c。

Gradio下载工具界面：

预设下载模块中预置了一些模型的下载链接和存放路径，首先选择一级分类和二级分类，然后选择预设文件列表（可多选），最后点击开始下载或生成终端下载代码。

Aria2c和Wget模块均可自定义下载链接和保存路径，对于多数下载链接可以直接识别文件名，无法识别的需要手工输入。

注意：

  1. 在工具内下载是有进度条的，但受网络影响，该进度条的刷新可能有延迟。如果想看完整的下载进度，建议生成终端下载代码后在新的终端内下载;
  2. 预设下载模块中的文件链接收集于互联网（如AI Models），不可商用！不可商用！不可商用！

4. 提交或删除文件/文件夹

提交文件或文件夹（终端内运行）：

# add 文件或文件夹的路径1 文件或文件夹的路径2 文件或文件夹的路径3 ...

# 例如提交comfyui、custom_nodes、venv文件夹
add comfyui custom_nodes venv

注意：可在资源管理器内，右键文件或文件夹，复制路径。

删除文件或文件夹（终端内运行，谨慎操作！）：

del 文件或文件夹的路径1 文件或文件夹的路径2 文件或文件夹的路径3 ...

5. 更新ComfyUI

终端内运行：update

6. 其他一些说明

预设下载中有的模型尽量不要提交到自己的仓库，预设链接是内网链接，下载很快。免费空间只有100+100GB，只要提交文件就会占用空间，后续即使删除文件，空间占用不会降低。需要释放仓库空间只能是删除仓库（注意下载备份重要文件！）。

venv生成方法

# 生成venv
python3 -m venv venv

# 激活venv
source venv/bin/activate

# 安装pytorch（注意cuda版本）
pip install torch==2.7.1 torchvision==0.22.1 torchaudio==2.7.1 --index-url https://download.pytorch.org/whl/cu128

# pip install torch==2.8.0 torchvision==0.23.0 torchaudio==2.8.0 --index-url https://download.pytorch.org/whl/cu129

# 安装dev版本的pytorch（注意cuda版本）
pip install -U torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu129

# 安装xformers（注意cuda版本，低版本pytorch需要手工指定xformers版本）
pip install xformers --index-url https://download.pytorch.org/whl/cu128

# pytorch版本 xformers版本对照表（仅供参考）
# torch==2.8.0 torchvision==0.23.0 torchaudio==2.8.0 xformers==0.0.32.post2
# torch==2.7.1 torchvision==0.22.1 torchaudio==2.7.1 xformers==0.0.31.post1
# torch==2.7.0 torchvision==0.22.0 torchaudio==2.7.0 xformers==0.0.30
# torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 xformers==0.0.29.post2
# torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 xformers==0.0.28.post3
# torch==2.5.0 torchvision==0.20.0 torchaudio==2.5.0 xformers==0.0.28.post2
# torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 xformers==0.0.28.post1
# torch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 xformers==0.0.27.post2
# torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 xformers==0.0.27
# torch==2.3.0 torchvision==0.18.0 torchaudio==2.3.0 xformers==0.0.26.post1
# torch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 xformers==0.0.26
# torch==2.2.1 torchvision==0.17.1 torchaudio==2.2.1 xformers==0.0.25
# torch==2.2.0 torchvision==0.17.0 torchaudio==2.2.0 xformers==0.0.24
# torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 xformers==0.0.23.post1
# torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 xformers==0.0.23
# torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 xformers==0.0.22.post7
# torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 无xformers
# torch==2.0.0 torchvision==0.15.1 torchaudio==2.0.1 无xformers

# 从源码安装xformers（pip安装的版本有问题时可尝试，低版本torch需要自己找低版本xformers源码）
git clone https://github.com/facebookresearch/xformers.git
cd xformers
git submodule update --init --recursive
python setup.py install
cd ..

# 安装flash-attn（具体版本得看插件要求）
pip install flash-attn==2.8.0.post2

# 安装SageAttention
git clone https://github.com/thu-ml/SageAttention.git
cd SageAttention 
export EXT_PARALLEL=4 NVCC_APPEND_FLAGS="--threads 8" MAX_JOBS=32 # parallel compiling (Optional)
python setup.py install
cd ..

安装ComfyUI及插件

# 下载ComfyUI
git clone https://github.com/comfyanonymous/ComfyUI.git /workspace/comfyui

# 下载常用插件
cd /workspace/custom_nodes
git clone https://github.com/1038lab/ComfyUI-RMBG
git clone https://github.com/7BEII/Comfyui_PDuse
git clone https://github.com/Acly/comfyui-inpaint-nodes
git clone https://github.com/AIFSH/ComfyUI-UVR5
git clone https://github.com/AInseven/ComfyUI-fastblend
git clone https://github.com/AIrjen/OneButtonPrompt
git clone https://github.com/badxprogramm/ComfyUI-GradientBlur
git clone https://github.com/BlenderNeko/ComfyUI_ADV_CLIP_emb
git clone https://github.com/bronkula/comfyui-fitsize
git clone https://github.com/catscandrive/comfyui-imagesubfolders
git clone https://github.com/chflame163/ComfyUI_FaceSimilarity
git clone https://github.com/chflame163/ComfyUI_LayerStyle
git clone https://github.com/chflame163/ComfyUI_LayerStyle_Advance
git clone https://github.com/chrisgoringe/cg-image-filter
git clone https://github.com/chrisgoringe/cg-use-everywhere
git clone https://github.com/christian-byrne/audio-separation-nodes-comfyui
git clone https://github.com/city96/ComfyUI-GGUF
git clone https://github.com/coreyryanhanson/ComfyQR
git clone https://github.com/crystian/comfyui-crystools
git clone https://github.com/cubiq/ComfyUI_essentials
git clone https://github.com/cubiq/ComfyUI_InstantID
git clone https://github.com/cubiq/ComfyUI_IPAdapter_plus
git clone https://github.com/cubiq/PuLID_ComfyUI
git clone https://github.com/CY-CHENYUE/ComfyUI-Janus-Pro
git clone https://github.com/CY-CHENYUE/ComfyUI-Redux-Prompt
git clone https://github.com/da2el-ai/D2-SavePSD-ComfyUI
git clone https://github.com/djbielejeski/a-person-mask-generator
git clone https://github.com/Dontdrunk/ComfyUI-DD-Translation
git clone https://github.com/dorpxam/ComfyUI-LTXVideoLoRA
git clone https://github.com/EllangoK/ComfyUI-post-processing-nodes
git clone https://github.com/evanspearman/ComfyMath
git clone https://github.com/EvilBT/ComfyUI_SLK_joy_caption_two
git clone https://github.com/Extraltodeus/Skimmed_CFG
git clone https://github.com/fairy-root/ComfyUI-OpenAI-FM
git clone https://github.com/fairy-root/Flux-Prompt-Generator
git clone https://github.com/Fannovel16/comfyui_controlnet_aux
git clone https://github.com/Fannovel16/ComfyUI-Frame-Interpolation
git clone https://github.com/Fannovel16/ComfyUI-Video-Matting
git clone https://github.com/FizzleDorf/ComfyUI_FizzNodes
git clone https://github.com/Flow-two/ComfyUI-WanStartEndFramesNative
git clone https://github.com/fpgaminer/joycaption_comfyui
git clone https://github.com/FunnyFinger/ComfyUi-RadarWeightNode
git clone https://github.com/glifxyz/ComfyUI-GlifNodes
git clone https://github.com/Goktug/ComfyUi_NNLatentUpscale
git clone https://github.com/Goktug/comfyui-saveimage-plus
git clone https://github.com/Gourieff/ComfyUI-ReActor
git clone https://github.com/GrailGreg/images_base64
git clone https://github.com/huchenlei/ComfyUI-IC-Light-Native
git clone https://github.com/huchenlei/ComfyUI-layerdiffuse
git clone https://github.com/huchenlei/ComfyUI-openpose-editor
git clone https://github.com/jags111/efficiency-nodes-comfyui
git clone https://github.com/jamesWalker55/comfyui-various
git clone https://github.com/jax-explorer/ComfyUI-VideoBasic
git clone https://github.com/john-mnz/ComfyUI-Inspyrenet-Rembg
git clone https://github.com/Jonseed/ComfyUI-Detail-Daemon
git clone https://github.com/kaibioinfo/ComfyUI_AdvancedRefluxControl
git clone https://github.com/kijai/ComfyUI-DepthAnythingV2
git clone https://github.com/kijai/ComfyUI-Florence2
git clone https://github.com/kijai/ComfyUI-GIMM-VFI
git clone https://github.com/kijai/ComfyUI-HunyuanVideoWrapper
git clone https://github.com/kijai/ComfyUI-IC-Light
git clone https://github.com/kijai/ComfyUI-KJNodes
git clone https://github.com/kijai/ComfyUI-Marigold
git clone https://github.com/kijai/ComfyUI-WanVideoWrapper
git clone https://github.com/Kosinkadink/ComfyUI-Advanced-ControlNet
git clone https://github.com/Kosinkadink/ComfyUI-AnimateDiff-Evolved
git clone https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite
git clone https://github.com/LAOGOU-666/Comfyui_LG_Tools
git clone https://github.com/Layer-norm/comfyui-lama-remover
git clone https://github.com/Lightricks/ComfyUI-LTXVideo
git clone https://github.com/lldacing/ComfyUI_Patches_ll
git clone https://github.com/lldacing/ComfyUI_PuLID_Flux_ll
git clone https://github.com/Loewen-Hob/rembg-comfyui-node-better
git clone https://github.com/lquesada/ComfyUI-Inpaint-CropAndStitch
git clone https://github.com/lrzjason/Comfyui-In-Context-Lora-Utils
git clone https://github.com/ltdrdata/ComfyUI-Impact-Pack
git clone https://github.com/ltdrdata/ComfyUI-Impact-Subpack
git clone https://github.com/ltdrdata/ComfyUI-Inspire-Pack
git clone https://github.com/ltdrdata/ComfyUI-Manager
git clone https://github.com/ltdrdata/was-node-suite-comfyui
git clone https://github.com/M1kep/ComfyLiterals
git clone https://github.com/melMass/comfy_mtb
git clone https://github.com/melMass/comfy_mtb
git clone https://github.com/mingsky-ai/ComfyUI-MingNodes
git clone https://github.com/neverbiasu/ComfyUI-SAM2
git clone https://github.com/nullquant/ComfyUI-BrushNet
git clone https://github.com/orssorbit/ComfyUI-wanBlockswap
git clone https://github.com/PressWagon/ComfyUI-StringsAndThings
git clone https://github.com/pythongosssss/ComfyUI-Custom-Scripts
git clone https://github.com/pythongosssss/ComfyUI-WD14-Tagger
git clone https://github.com/raindrop313/ComfyUI-WanVideoStartEndFrames
git clone https://github.com/shadowcz007/comfyui-mixlab-nodes
git clone https://github.com/Shakker-Labs/ComfyUI-IPAdapter-Flux
git clone https://github.com/sipherxyz/comfyui-art-venture
git clone https://github.com/smthemex/ComfyUI_Pops
git clone https://github.com/smthemex/ComfyUI_Sonic
git clone https://github.com/smthemex/ComfyUI_YuE
git clone https://github.com/ssitu/ComfyUI_UltimateSDUpscale
git clone https://github.com/Stability-AI/stability-ComfyUI-nodes
git clone https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes
git clone https://github.com/talesofai/comfyui-browser
git clone https://github.com/TFL-TFL/ComfyUI_Text_Translation
git clone https://github.com/TheMistoAI/ComfyUI-Anyline
git clone https://github.com/Tropfchen/ComfyUI-Embedding_Picker
git clone https://github.com/TTPlanetPig/Comfyui_TTP_Toolset
git clone https://github.com/un-seen/comfyui-tensorops
git clone https://github.com/VAST-AI-Research/ComfyUI-Tripo
git clone https://github.com/WaveSpeedAI/wavespeed-comfyui
git clone https://github.com/welltop-cn/ComfyUI-TeaCache
git clone https://github.com/xhiroga/ComfyUI-FramePackWrapper_PlusOne
git clone https://github.com/XLabs-AI/x-flux-comfyui
git clone https://github.com/yolain/ComfyUI-Easy-Use

# 批量安装插件依赖
find /workspace/custom_nodes -name requirements.txt -exec pip install -r {} \;

# 重新安装ComfyUI依赖
pip install -r /workspace/comfyui/requirements.txt

# 链接插件目录到ComfyUI目录
rm -r /workspace/comfyui/custom_nodes
ln -s /workspace/custom_nodes /workspace/comfyui/

# 启动ComfyUI看看还缺什么依赖或哪些依赖存在冲突，手工处理一下！！！
/workspace/venv/bin/python /workspace/comfyui/main.py --listen --auto-launch

其他插件：

# 启动较慢插件
git clone https://github.com/billwuhao/ComfyUI_PortraitTools

# UI类插件
git clone https://github.com/rgthree/rgthree-comfy

# 会生成独立按钮的插件
git clone https://github.com/ainewsto/Comfyui_Comfly
git clone https://github.com/bytedance/comfyui-lumi-batcher

# python3.12暂不支持的插件
git clone https://github.com/kijai/ComfyUI-SUPIR
git clone https://github.com/chflame163/ComfyUI_WordCloud
git clone https://github.com/ZHO-ZHO-ZHO/ComfyUI-YoloWorld-EfficientSAM
git clone https://github.com/filliptm/ComfyUI_Fill-Nodes
git clone https://github.com/AlekPet/ComfyUI_Custom_Nodes_AlekPet # 部分支持

ComfyUI源码修改

准备模型下载链接

为了实现自动下载模型，首先需要手工收集模型的下载链接，并保存到一个文件里。这里还需要指定模型的下载路径，建议使用json格式。

注意：

  1. checkpoints、clip、clip_vision、controlnet、diffusion_models、loras、text_encoders、upscale_models、vae文件夹是由ComfyUI底层函数搜索的，可以通过hook源码实现自动下载。

  2. 对于插件来说，得具体看插件是如何查找相关模型的，如果是调用ComfyUI底层函数，那该方法同样有效，否则无效。

示例：

{
    "checkpoints": {
        "hunyuan_dit_1.0.safetensors": "https://cnb.cool/ai-models/comfyanonymous/hunyuan_dit_comfyui/-/lfs/60ee5b765b40575c23a82cb4fec8e51e28157391e94397abf40d9546e81e1c58",
        "hunyuan_dit_1.1.safetensors": "https://cnb.cool/ai-models/comfyanonymous/hunyuan_dit_comfyui/-/lfs/26ab130e7bb515e976aa6d8fac8c3f11caef986c13140447ba102a4b1e675245",
        "hunyuan_dit_1.2.safetensors": "https://cnb.cool/ai-models/comfyanonymous/hunyuan_dit_comfyui/-/lfs/4fb84f84079cda457d171b3c6b15d1be95b5a3e5d9825703951a99ddf92d1787"
    },
    "clip_vision": {
        "clip_vision_g.safetensors": "https://cnb.cool/ai-models/comfyanonymous/clip_vision_g/-/lfs/9908329b3ead722a693ea400fab1d7c9ec91d6736fd194a94d20d793457f9c2e"
    }
}

ComfyUI源码定位及修改

注意：这里主要是介绍修改思路/原理，本仓库中的实际代码可能有所不同，是针对实际情况优化修改的（如模型选项去重等，这里不再介绍）。

folder_paths.py

将预设模型添加到节点选项中

原始代码函数定位：

def recursive_search(directory: str, excluded_dir_names: list[str] | None=None) -> tuple[list[str], dict[str, float]]:
    if not os.path.isdir(directory):
        return [], {}

    if excluded_dir_names is None:
        excluded_dir_names = []

    result = []
    dirs = {}

    # Attempt to add the initial directory to dirs with error handling
    try:
        dirs[directory] = os.path.getmtime(directory)
    except FileNotFoundError:
        logging.warning(f"Warning: Unable to access {directory}. Skipping this path.")

    logging.debug("recursive file list on directory {}".format(directory))
    dirpath: str
    subdirs: list[str]
    filenames: list[str]

    for dirpath, subdirs, filenames in os.walk(directory, followlinks=True, topdown=True):
        subdirs[:] = [d for d in subdirs if d not in excluded_dir_names]
        for file_name in filenames:
            try:
                relative_path = os.path.relpath(os.path.join(dirpath, file_name), directory)
                result.append(relative_path)
            except:
                logging.warning(f"Warning: Unable to access {file_name}. Skipping this file.")
                continue

        for d in subdirs:
            path: str = os.path.join(dirpath, d)
            try:
                dirs[path] = os.path.getmtime(path)
            except FileNotFoundError:
                logging.warning(f"Warning: Unable to access {path}. Skipping this path.")
                continue

    logging.debug("found {} files".format(len(result)))
    return result, dirs

修改为：

# 导入json模块用来加载模型下载链接文件，导入subprocess模块用于后续调用aria2c下载模型
import json, subprocess

# 假设存放模型下载链接的文件是/workspace/source.json
source_file = "/workspace/source.json"
all_file_dict = {}
if os.path.exists(source_file):
    with open(source_file, "r", encoding="utf-8") as file:
        all_file_dict = json.load(file)

def recursive_search(directory: str, excluded_dir_names: list[str] | None=None) -> tuple[list[str], dict[str, float]]:
    if not os.path.isdir(directory):
        return [], {}

    if excluded_dir_names is None:
        excluded_dir_names = []

    result = []
    dirs = {}

    # Attempt to add the initial directory to dirs with error handling
    try:
        dirs[directory] = os.path.getmtime(directory)
    except FileNotFoundError:
        logging.warning(f"Warning: Unable to access {directory}. Skipping this path.")

    logging.debug("recursive file list on directory {}".format(directory))
    dirpath: str
    subdirs: list[str]
    filenames: list[str]

    for dirpath, subdirs, filenames in os.walk(directory, followlinks=True, topdown=True):
        subdirs[:] = [d for d in subdirs if d not in excluded_dir_names]
        for file_name in filenames:
            try:
                relative_path = os.path.relpath(os.path.join(dirpath, file_name), directory)
                result.append(relative_path)
            except:
                logging.warning(f"Warning: Unable to access {file_name}. Skipping this file.")
                continue

        for d in subdirs:
            path: str = os.path.join(dirpath, d)
            try:
                dirs[path] = os.path.getmtime(path)
            except FileNotFoundError:
                logging.warning(f"Warning: Unable to access {path}. Skipping this path.")
                continue

    # 将all_file_dict中的模型添加到节点选项中，这里注意models文件夹的实际绝对位置
    if "/workspace/comfyui/models/" in directory:
        folder = directory.split("/workspace/comfyui/models/")[-1]
        if folder in all_file_dict:
            for file in all_file_dict[folder]:
                # 跳过已经存在的文件选项
                if file not in result:
                    result.append(file)

    logging.debug("found {} files".format(len(result)))
    return result, dirs

当文件不存在时自动下载

原始代码函数定位：

def get_full_path_or_raise(folder_name: str, filename: str) -> str:
    """
    Get the full path of a file in a folder, has to be a file
    """
    full_path = get_full_path(folder_name, filename)

    if full_path is None:
        raise FileNotFoundError(f"Model in folder '{folder_name}' with filename '{filename}' not found.")
    return full_path

修改为：

def get_full_path_or_raise(folder_name: str, filename: str) -> str:
    """
    Get the full path of a file in a folder, has to be a file
    """
    full_path = get_full_path(folder_name, filename)

    # 当没有找到模型时，尝试下载，注意models_path的位置，本仓库的.cnb.yml中将/models挂载在云节点缓存里
    if full_path is None:
        models_path = "/models"
        url = ""
        if folder_name in all_file_dict:
            file_dict = all_file_dict[folder_name]
            if filename in file_dict:
                url = file_dict[filename]
                file_path = os.path.join(models_path, folder_name, filename)
                folder_path = os.path.dirname(file_path)
                file_name = os.path.basename(file_path)
                os.makedirs(folder_path, exist_ok=True)
                subprocess.run(f'aria2c -x 4 -s 4 -c -d "{folder_path}" -o "{file_name}" "{url}"', shell=True)
                if os.path.isfile(file_path):
                    full_path = file_path

    if full_path is None:
        raise FileNotFoundError(f"Model in folder '{folder_name}' with filename '{filename}' not found.")
    return full_path

app/model_manager.py

将预设模型添加到UI界面选项中

将预设模型添加到UI界面选项中，主要是避免UI界面弹出模型缺失的提示

原始代码函数定位：

    def recursive_search_models_(self, directory: str, pathIndex: int) -> tuple[list[str], dict[str, float], float]:
        if not os.path.isdir(directory):
            return [], {}, time.perf_counter()

        excluded_dir_names = [".git"]
        # TODO use settings
        include_hidden_files = False

        result: list[str] = []
        dirs: dict[str, float] = {}

        for dirpath, subdirs, filenames in os.walk(directory, followlinks=True, topdown=True):
            subdirs[:] = [d for d in subdirs if d not in excluded_dir_names]
            if not include_hidden_files:
                subdirs[:] = [d for d in subdirs if not d.startswith(".")]
                filenames = [f for f in filenames if not f.startswith(".")]

            filenames = filter_files_extensions(filenames, folder_paths.supported_pt_extensions)

            for file_name in filenames:
                try:
                    full_path = os.path.join(dirpath, file_name)
                    relative_path = os.path.relpath(full_path, directory)

                    # Get file metadata
                    file_info = {
                        "name": relative_path,
                        "pathIndex": pathIndex,
                        "modified": os.path.getmtime(full_path),  # Add modification time
                        "created": os.path.getctime(full_path),   # Add creation time
                        "size": os.path.getsize(full_path)        # Add file size
                    }
                    result.append(file_info)

                except Exception as e:
                    logging.warning(f"Warning: Unable to access {file_name}. Error: {e}. Skipping this file.")
                    continue

            for d in subdirs:
                path: str = os.path.join(dirpath, d)
                try:
                    dirs[path] = os.path.getmtime(path)
                except FileNotFoundError:
                    logging.warning(f"Warning: Unable to access {path}. Skipping this path.")
                    continue

        return result, dirs, time.perf_counter()

修改为：

    def recursive_search_models_(self, directory: str, pathIndex: int) -> tuple[list[str], dict[str, float], float]:
        if not os.path.isdir(directory):
            return [], {}, time.perf_counter()

        excluded_dir_names = [".git"]
        # TODO use settings
        include_hidden_files = False

        result: list[str] = []
        dirs: dict[str, float] = {}

        for dirpath, subdirs, filenames in os.walk(directory, followlinks=True, topdown=True):
            subdirs[:] = [d for d in subdirs if d not in excluded_dir_names]
            if not include_hidden_files:
                subdirs[:] = [d for d in subdirs if not d.startswith(".")]
                filenames = [f for f in filenames if not f.startswith(".")]

            filenames = filter_files_extensions(filenames, folder_paths.supported_pt_extensions)

            for file_name in filenames:
                try:
                    full_path = os.path.join(dirpath, file_name)
                    relative_path = os.path.relpath(full_path, directory)

                    # Get file metadata
                    file_info = {
                        "name": relative_path,
                        "pathIndex": pathIndex,
                        "modified": os.path.getmtime(full_path),  # Add modification time
                        "created": os.path.getctime(full_path),   # Add creation time
                        "size": os.path.getsize(full_path)        # Add file size
                    }
                    result.append(file_info)

                except Exception as e:
                    logging.warning(f"Warning: Unable to access {file_name}. Error: {e}. Skipping this file.")
                    continue

            for d in subdirs:
                path: str = os.path.join(dirpath, d)
                try:
                    dirs[path] = os.path.getmtime(path)
                except FileNotFoundError:
                    logging.warning(f"Warning: Unable to access {path}. Skipping this path.")
                    continue


        # 添加预设模型，这里同样注意models文件夹的实际绝对路径
        if "/workspace/comfyui/models/" in directory:
            folder = directory.split("/workspace/comfyui/models/")[-1]

            # 从folder_paths.py中导入all_file_dict
            from folder_paths import all_file_dict

            if folder in all_file_dict:
                for file in all_file_dict[folder]:
                    found = False
                    for file_info in result:
                        if file == file_info["name"]:
                            found = True
                    if not found:
                        # 注意这里的file_info，实际上应包括pathIndex、modified、created、size四个参数，这里设置为0就可以避免弹出缺少模型的提示了
                        # 但都设置为0，在UI界面中，点击模型库模块里的模型，可能没有反应
                        file_info = {
                            "name": file,
                            "pathIndex": 0,
                            "modified": 0,
                            "created": 0,
                            "size": 0
                        }
                        result.append(file_info)
                        

        return result, dirs, time.perf_counter()

extra_model_paths.yaml

该文件主要是添加额外的模型搜索路径，这里添加/models文件夹，是为了把自动下载的模型缓存在云节点上。需要搭配.cnb.yml文件设置。

comfyui:
     base_path: /models/

     checkpoints: checkpoints/
     clip: clip/
     clip_vision: clip_vision/
     configs: configs/
     controlnet: controlnet/
     diffusion_models: |
                  diffusion_models
                  unet
     embeddings: embeddings/
     loras: loras/
     upscale_models: upscale_models/
     vae: vae/

.cnb.yml

通过volumes设置缓存文件夹/models

$:
  vscode:
    - docker:
        image: docker.cnb.cool/cnb-xu/docker/comfyui:latest
        volumes:
          - /models

About

云原生构建的ComfyUI运行环境 主要研究： 1. 快速构建自定义运行环境 2. 初始化阶段自定义模型下载 3. 基于Gradio的图形化下载工具 4. ComfyUI底层自动检测/下载预设模型 5. 固定镜像环境下自定义内容（插件、依赖、模型、设置等）的保存（提交）

984.04 MiB
0 forks0 stars1 branches0 TagREADME
Language
CSV34.9%
Python23.6%
JavaScript18.8%
XML7.3%
Others15.4%
© 2025 Tencent, Inc. Supported by Tencent Cloud
About Us
Privacy Policy
Terms of Service
SLA
API
Docs
Showcase
Feedback
35/F,Tencent Building,Kejizhongyi Avenue,Nanshan District,Shenzhen
京ICP备11018762号-111
粤公网安备44030002006058号粤公网安备44030002006058号