The backend crashes when loading the model with mmap.

[304523475]

Custom Node Testing

• I have tried disabling custom nodes and the issue persists (see how to disable custom nodes if you need help)

Expected Behavior

FETCH ComfyRegistry Data: 10/135
Exception in callback _ProactorBasePipeTransport._call_connection_lost(None)
handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)>
Traceback (most recent call last):
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88, in _run
self._context.run(self._callback, *self._args)
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\proactor_events.py", line 165, in _call_connection_lost
self._sock.shutdown(socket.SHUT_RDWR)
ConnectionResetError: [WinError 10054] 远程主机强迫关闭了一个现有的连接。
Exception in callback _ProactorBasePipeTransport._call_connection_lost(None)
handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)>
Traceback (most recent call last):
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88, in _run
self._context.run(self._callback, *self._args)
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\proactor_events.py", line 165, in _call_connection_lost
self._sock.shutdown(socket.SHUT_RDWR)
ConnectionResetError: [WinError 10054] 远程主机强迫关闭了一个现有的连接。
FETCH ComfyRegistry Data: 15/135
FETCH ComfyRegistry Data: 20/135
FETCH ComfyRegistry Data: 25/135
FETCH ComfyRegistry Data: 30/135
FETCH ComfyRegistry Data: 35/135
FETCH ComfyRegistry Data: 40/135
FETCH ComfyRegistry Data: 45/135
got prompt
Failed to validate prompt for output 704:

• (prompt):
  • Required input is missing: images
• VHS_VideoCombine 704:
  • Required input is missing: images
    Output will be ignored
    FETCH ComfyRegistry Data: 50/135
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.59GB (自动模式: 总显存=24.00GB, 已用=1.99GB)
    [DynamicRAMCache] Switched mode: CLASSIC -> RAM_PRESSURE (Headroom: 4.0GB)
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.54GB (自动模式: 总显存=24.00GB, 已用=1.94GB)
    FETCH ComfyRegistry Data: 55/135
    FETCH ComfyRegistry Data: 60/135
    FETCH ComfyRegistry Data: 65/135
    FETCH ComfyRegistry Data: 70/135
    FETCH ComfyRegistry Data: 75/135
    FETCH ComfyRegistry Data: 80/135
    FETCH ComfyRegistry Data: 85/135
    CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.53GB (自动模式: 总显存=24.00GB, 已用=1.93GB)
    FETCH ComfyRegistry Data: 90/135
    Found quantization metadata version 1
    Detected mixed precision quantization
    Using mixed precision operations
    model weight dtype torch.bfloat16, manual cast: torch.bfloat16
    model_type FLUX
    VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
    no CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
    FETCH ComfyRegistry Data: 95/135
    Requested to load LTXAVTEModel_
    FETCH ComfyRegistry Data: 100/135
    loaded partially; 18774.56 MB usable, 18212.06 MB loaded, 7228.54 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
    loaded partially; 18474.69 MB usable, 17912.19 MB loaded, 7528.57 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
    Requested to load VideoVAE
    FETCH ComfyRegistry Data: 105/135
    loaded completely; 19823.23 MB usable, 1384.94 MB loaded, full load: True
    FETCH ComfyRegistry Data: 110/135
    Windows fatal exception: access violation

Stack (most recent call first):
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\storage.py", line 466 in getitem
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 136 in load_torch_file
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_lt_audio.py", line 30 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\latest_io.py", line 1764 in EXECUTE_NORMALIZED
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\internal_init_.py", line 149 in wrapped_func
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 296 in process_inputs
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 308 in _async_map_node_over_list
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 334 in get_output_data
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 525 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 758 in execute_async
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88 in _run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 1986 in _run_once
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 641 in run_forever
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\windows_events.py", line 322 in run_forever
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 674 in run_until_complete
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 118 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 194 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 702 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\main.py", line 276 in prompt_worker
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1012 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1075 in _bootstrap_inner
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1032 in _bootstrap

H:\ai\comfyui2\ComfyUI_windows_portable>pause
请按任意键继续. . .

Actual Behavior

我的本地显卡为3090，64g内存，当我跑ltx2.3时，偶尔加载成功可以跑通工作流，但是经常是会出现加载模型时后台自动崩溃，比如跑完ltx2.3的工作流时，再换个其他的工作流跑时就一定会出现后台崩溃。以下是我问claud的错误回复：

这个 bug 在 comfy_aimdo 模块里，让我看看它是 Python 代码还是编译的二进制文件：

Ran command: Get-ChildItem -Path "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfy_aimdo" -Recurse | Select-Object FullName, Extension, Length | Format-Table -AutoSize
Ran command: Get-ChildItem -Path "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfy_aimdo" -Recurse | ForEach-Object { $_.FullName + " | " + $_.Extension + " | " + $_.Length }
Listed directory comfy_aimdo
Viewed model_mmap.py:1-48

这就是答案了：

没办法修复，因为 bug 在编译好的 DLL 里

comfy_aimdo/
├── aimdo.dll          ← 223KB 编译的二进制文件（C/C++ 写的）
├── model_mmap.py      ← Python 封装层，只是调用 DLL
└── ...

model_mmap.py 只是一层薄封装，真正的 mmap 逻辑在 aimdo.dll 里：

self.state = lib.model_mmap_allocate(filepath_bytes)  # → 调用 DLL
return lib.model_mmap_get(self.state)                  # → 调用 DLL

• aimdo.dll 是 ComfyUI 官方团队用 C/C++ 编译的闭源二进制文件
• 我们看不到源码，改不了
• access violation 就发生在 DLL 返回的内存地址被 PyTorch 访问时

所以只有两个选择

方案	可行性
修复 aimdo.dll	❌ 闭源二进制，无法修改，只能等官方更新
绕过它（--disable-mmap）	✅ 已帮你做好，代价极小

用 --disable-mmap 后模型文件改用标准 safetensors.safe_open() 方式加载，加载速度增加几秒，但推理速度完全不受影响。这是目前唯一可行的方案。

Steps to Reproduce

当我跑ltx2.3时，偶尔加载成功可以跑通工作流，但是经常是会出现加载模型时后台自动崩溃，比如跑完ltx2.3的工作流时，再换个其他的工作流跑时如z image就一定会出现后台崩溃。

Debug Logs

FETCH ComfyRegistry Data: 10/135
Exception in callback _ProactorBasePipeTransport._call_connection_lost(None)
handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)>
Traceback (most recent call last):
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88, in _run
    self._context.run(self._callback, *self._args)
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\proactor_events.py", line 165, in _call_connection_lost
    self._sock.shutdown(socket.SHUT_RDWR)
ConnectionResetError: [WinError 10054] 远程主机强迫关闭了一个现有的连接。
Exception in callback _ProactorBasePipeTransport._call_connection_lost(None)
handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)>
Traceback (most recent call last):
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88, in _run
    self._context.run(self._callback, *self._args)
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\proactor_events.py", line 165, in _call_connection_lost
    self._sock.shutdown(socket.SHUT_RDWR)
ConnectionResetError: [WinError 10054] 远程主机强迫关闭了一个现有的连接。
FETCH ComfyRegistry Data: 15/135
FETCH ComfyRegistry Data: 20/135
FETCH ComfyRegistry Data: 25/135
FETCH ComfyRegistry Data: 30/135
FETCH ComfyRegistry Data: 35/135
FETCH ComfyRegistry Data: 40/135
FETCH ComfyRegistry Data: 45/135
got prompt
Failed to validate prompt for output 704:
* (prompt):
  - Required input is missing: images
* VHS_VideoCombine 704:
  - Required input is missing: images
Output will be ignored
FETCH ComfyRegistry Data: 50/135
[ReservedVRAM]执行前置GPU显存清理...
[ReservedVRAM]GPU显存清理完成
[ReservedVRAM]set EXTRA_RESERVED_VRAM=2.59GB (自动模式: 总显存=24.00GB, 已用=1.99GB)
[DynamicRAMCache] Switched mode: CLASSIC -> RAM_PRESSURE (Headroom: 4.0GB)
[ReservedVRAM]执行前置GPU显存清理...
[ReservedVRAM]GPU显存清理完成
[ReservedVRAM]set EXTRA_RESERVED_VRAM=2.54GB (自动模式: 总显存=24.00GB, 已用=1.94GB)
FETCH ComfyRegistry Data: 55/135
FETCH ComfyRegistry Data: 60/135
FETCH ComfyRegistry Data: 65/135
FETCH ComfyRegistry Data: 70/135
FETCH ComfyRegistry Data: 75/135
FETCH ComfyRegistry Data: 80/135
FETCH ComfyRegistry Data: 85/135
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
[ReservedVRAM]执行前置GPU显存清理...
[ReservedVRAM]GPU显存清理完成
[ReservedVRAM]set EXTRA_RESERVED_VRAM=2.53GB (自动模式: 总显存=24.00GB, 已用=1.93GB)
FETCH ComfyRegistry Data: 90/135
Found quantization metadata version 1
Detected mixed precision quantization
Using mixed precision operations
model weight dtype torch.bfloat16, manual cast: torch.bfloat16
model_type FLUX
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
no CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
FETCH ComfyRegistry Data: 95/135
Requested to load LTXAVTEModel_
FETCH ComfyRegistry Data: 100/135
loaded partially; 18774.56 MB usable, 18212.06 MB loaded, 7228.54 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
loaded partially; 18474.69 MB usable, 17912.19 MB loaded, 7528.57 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
Requested to load VideoVAE
FETCH ComfyRegistry Data: 105/135
loaded completely; 19823.23 MB usable, 1384.94 MB loaded, full load: True
FETCH ComfyRegistry Data: 110/135
Windows fatal exception: access violation

Stack (most recent call first):
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\storage.py", line 466 in __getitem__
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 136 in load_torch_file
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_lt_audio.py", line 30 in execute
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\latest\_io.py", line 1764 in EXECUTE_NORMALIZED
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\internal\__init__.py", line 149 in wrapped_func
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 296 in process_inputs
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 308 in _async_map_node_over_list
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 334 in get_output_data
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 525 in execute
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 758 in execute_async
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88 in _run
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 1986 in _run_once
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 641 in run_forever
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\windows_events.py", line 322 in run_forever
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 674 in run_until_complete
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 118 in run
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 194 in run
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 702 in execute
  File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\main.py", line 276 in prompt_worker
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1012 in run
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1075 in _bootstrap_inner
  File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1032 in _bootstrap

H:\ai\comfyui2\ComfyUI_windows_portable>pause
请按任意键继续. . .

Other

No response

[rattus128]

Can you paste your startup log? It should show why aimdo failed to probe and may have more information. You seem to be in full fallback to non-dynamic vram yet its using the dyn vram loader so its a bit weird.

[304523475]

Can you paste your startup log? It should show why aimdo failed to probe and may have more information. You seem to be in full fallback to non-dynamic vram yet its using the dyn vram loader so its a bit weird.能否粘贴一下启动日志？它应该能显示 aimdo 探测失败的原因，并且可能包含更多信息。你似乎完全回退到了非动态显存，但它却在使用动态显存加载器，这有点奇怪。

下面是完全官方的启动命令：

H:\ai\comfyui2\ComfyUI_windows_portable>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --port 3721
Adding extra search path checkpoints H:\ai\ComfyUI_windows_portable\ComfyUI\models\checkpoints
Adding extra search path clip H:\ai\ComfyUI_windows_portable\ComfyUI\models\clip
Adding extra search path clip_vision H:\ai\ComfyUI_windows_portable\ComfyUI\models\clip_vision
Adding extra search path configs H:\ai\ComfyUI_windows_portable\ComfyUI\models\configs
Adding extra search path controlnet H:\ai\ComfyUI_windows_portable\ComfyUI\models\controlnet
Adding extra search path diffusion_models H:\ai\ComfyUI_windows_portable\ComfyUI\models\diffusion_models
Adding extra search path diffusion_models H:\ai\ComfyUI_windows_portable\ComfyUI\models\unet
Adding extra search path embeddings H:\ai\ComfyUI_windows_portable\ComfyUI\models\embeddings
Adding extra search path loras H:\ai\ComfyUI_windows_portable\ComfyUI\models\loras
Adding extra search path upscale_models H:\ai\ComfyUI_windows_portable\ComfyUI\models\upscale_models
Adding extra search path vae H:\ai\ComfyUI_windows_portable\ComfyUI\models\vae
Adding extra search path text_encoders H:\ai\ComfyUI_windows_portable\ComfyUI\models\text_encoders
Adding extra search path animatediff_models H:\ai\ComfyUI_windows_portable\ComfyUI\models\animatediff_models
Adding extra search path detection H:\ai\ComfyUI_windows_portable\ComfyUI\models\detection
Adding extra search path florence2 H:\ai\ComfyUI_windows_portable\ComfyUI\models\florence2
Adding extra search path yolo H:\ai\ComfyUI_windows_portable\ComfyUI\models\yolo
Adding extra search path model_patches H:\ai\ComfyUI_windows_portable\ComfyUI\models\model_patches
Adding extra search path latent_upscale_models H:\ai\ComfyUI_windows_portable\ComfyUI\models\latent_upscale_models
[START] Security scan
[DONE] Security scan

ComfyUI-Manager: installing dependencies done.

** ComfyUI startup time: 2026-03-30 00:06:10.315
** Platform: Windows
** Python version: 3.12.7 (tags/v3.12.7:0b05ead, Oct 1 2024, 03:06:41) [MSC v.1941 64 bit (AMD64)]
** Python executable: H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\python.exe
** ComfyUI Path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI
** ComfyUI Base Folder Path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI
** User directory: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user
** ComfyUI-Manager config path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user__manager\config.ini
** Log path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user\comfyui.log

Prestartup times for custom nodes:
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use
2.8 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager

H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\cuda_init_.py:61: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations.
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Checkpoint files will always be loaded safely.
Total VRAM 24576 MB, total RAM 65322 MB
pytorch version: 2.7.1+cu128
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29394.0
Using pytorch attention
Unsupported Pytorch detected. DynamicVRAM support requires Pytorch version 2.8 or later. Falling back to legacy ModelPatcher. VRAM estimates may be unreliable especially on Windows
Python version: 3.12.7 (tags/v3.12.7:0b05ead, Oct 1 2024, 03:06:41) [MSC v.1941 64 bit (AMD64)]
ComfyUI version: 0.18.2
comfy-aimdo version: 0.2.12
comfy-kitchen version: 0.2.8
ComfyUI frontend version: 1.41.21
[Prompt Server] web root: H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfyui_frontend_package\static
Asset seeder disabled
[Crystools INFO] Crystools version: 1.27.4
[Crystools INFO] Platform release: 11
[Crystools INFO] JETSON: Not detected.
[Crystools INFO] CPU: 12th Gen Intel(R) Core(TM) i7-12700K - Arch: AMD64 - OS: Windows 11
[Crystools INFO] pynvml (NVIDIA) initialized.
[Crystools INFO] GPU/s:
[Crystools INFO] 0) NVIDIA GeForce RTX 3090
[Crystools INFO] NVIDIA Driver: 595.79
[ComfyUI-Easy-Use] server: v1.3.6 Loaded
[ComfyUI-Easy-Use] web root: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use\web_version/v2 Loaded

Loading: ComfyUI-Impact-Pack (V8.28.2)

[Impact Pack] Wildcard total size (0.00 MB) is within cache limit (50.00 MB). Using full cache mode.
[Impact Pack] Wildcards loading done.

Loading: ComfyUI-Impact-Subpack (V1.3.5)

[Impact Pack/Subpack] Using folder_paths to determine whitelist path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Impact-Subpack\model-whitelist.txt
[Impact Pack/Subpack] Ensured whitelist directory exists: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Impact-Subpack
[Impact Pack/Subpack] Loaded 0 model(s) from whitelist: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Impact-Subpack\model-whitelist.txt
[Impact Subpack] ultralytics_bbox: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\models\ultralytics\bbox
[Impact Subpack] ultralytics_segm: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\models\ultralytics\segm

Loading: ComfyUI-Manager (V3.39)

[ComfyUI-Manager] network_mode: public
[ComfyUI-Manager] ComfyUI per-queue preview override detected (PR #11261). Manager's preview method feature is disabled. Use ComfyUI's --preview-method CLI option or 'Settings > Execution > Live preview method'.

ComfyUI Revision: 1500 [173e1aa] *DETACHED | Released on '2026-03-26'

⚡ SeedVR2 optimizations check: Flash Attention ✅ | Triton ✅
📊 Initial CUDA memory: 22.72GB free / 24.00GB total
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json

Comfyroll Studio v1.76 : 175 Nodes Loaded

** For changes, please see patch notes at https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes/blob/main/Patch_Notes.md
** For help, please see the wiki at https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes/wiki

[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
[H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ckpts path: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\ckpts
[H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using symlinks: False
[H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ort providers: ['CUDAExecutionProvider', 'DirectMLExecutionProvider', 'OpenVINOExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider', 'CoreMLExecutionProvider']
H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\node_wrappers\dwpose.py:26: UserWarning: DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly
warnings.warn("DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly")
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[tinyterraNodes] Loaded
[llama-cpp_vlm] Model cleanup hook applied!
✨提示词小助手 V2.0.4 已启动
'sox' 不是内部或外部命令，也不是可运行的程序
或批处理文件。
SoX could not be found!

If you do not have SoX, proceed here:
 - - - http://sox.sourceforge.net/ - - -

If you do (or think that you should) have SoX, double-check your
path variables.

Using pytorch attention
(RES4LYF) Init
(RES4LYF) Importing beta samplers.
(RES4LYF) Importing legacy samplers.

[rgthree-comfy] Loaded 48 fantastic nodes. 🎉

[rgthree-comfy] ComfyUI's new Node 2.0 rendering may be incompatible with some rgthree-comfy nodes and features, breaking some rendering as well as losing the ability to access a node's properties (a vital part of many nodes). It also appears to run MUCH more slowly spiking CPU usage and causing jankiness and unresponsiveness, especially with large workflows. Personally I am not planning to use the new Nodes 2.0 and, unfortunately, am not able to invest the time to investigate and overhaul rgthree-comfy where needed. If you have issues when Nodes 2.0 is enabled, I'd urge you to switch it off as well and join me in hoping ComfyUI is not planning to deprecate the existing, stable canvas rendering all together.

Import times for custom nodes:
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\websocket_image_save.py
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\Comfyui-PainterAudioCut
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_Video_Dynamic_Crop
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_novel_loader_light
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-qwenmultiangle
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\dynamic-ramcache
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_ttp_toolset
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\seedvarianceenhancer
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-inpaint-cropandstitch
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\reservedvram
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\Z-Image-Turbo-Lora-Stack-V4
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\LTX-2-3-LipSync
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_EncryptedWorkflow
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\AnySwitch
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\erosdiffusion-eulerflowmatchingdiscretescheduler
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-impact-subpack
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\LanPaint
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-DyPE
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_essentials
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-custom-scripts
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-MelBandRoFormer
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_VNCCS_Utils_JPAI
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-KJNodes
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\llama-cpp_vllm
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-LTXVideo
0.0 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes
0.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\prompt-assistant
0.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux
0.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_nvidia_rtx_nodes
0.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_tinyterranodes
0.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-VideoHelperSuite
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-impact-pack
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-SeedVR2_VideoUpscaler
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Crystools
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-detail-daemon
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_LayerStyle
0.2 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-advancedliveportrait
0.3 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\RES4LYF
0.4 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-BiRefNet-Hugo
0.4 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager
1.1 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_lg_tools
1.4 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_LayerStyle_Advance
1.7 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use
1.7 seconds: H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\custom_nodes\qwen3-tts

Context impl SQLiteImpl.
Will assume non-transactional DDL.
Starting server

To see the GUI go to: http://127.0.0.1:3721
FETCH ComfyRegistry Data: 5/135
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui/components/buttonGroup.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /extensions/core/clipspace.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /extensions/core/groupNode.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /extensions/core/widgetInputs.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui/components/button.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
Exception in callback _ProactorBasePipeTransport._call_connection_lost(None)
handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)>
Traceback (most recent call last):
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88, in _run
self._context.run(self._callback, *self._args)
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\proactor_events.py", line 165, in _call_connection_lost
self._sock.shutdown(socket.SHUT_RDWR)
ConnectionResetError: [WinError 10054] 远程主机强迫关闭了一个现有的连接。
FETCH ComfyRegistry Data: 10/135
FETCH ComfyRegistry Data: 15/135
FETCH ComfyRegistry Data: 20/135
FETCH ComfyRegistry Data: 25/135
FETCH ComfyRegistry Data: 30/135
FETCH ComfyRegistry Data: 35/135
FETCH ComfyRegistry Data: 40/135
FETCH ComfyRegistry Data: 45/135
got prompt
Failed to validate prompt for output 704:

• (prompt):
  • Required input is missing: images
• VHS_VideoCombine 704:
  • Required input is missing: images
    Output will be ignored
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.44GB (自动模式: 总显存=24.00GB, 已用=1.84GB)
    [DynamicRAMCache] Switched mode: CLASSIC -> RAM_PRESSURE (Headroom: 4.0GB)
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.44GB (自动模式: 总显存=24.00GB, 已用=1.84GB)
    FETCH ComfyRegistry Data: 50/135
    FETCH ComfyRegistry Data: 55/135
    FETCH ComfyRegistry Data: 60/135
    CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
    [ReservedVRAM]执行前置GPU显存清理...
    [ReservedVRAM]GPU显存清理完成
    [ReservedVRAM]set EXTRA_RESERVED_VRAM=2.43GB (自动模式: 总显存=24.00GB, 已用=1.83GB)
    Found quantization metadata version 1
    Detected mixed precision quantization
    Using mixed precision operations
    model weight dtype torch.bfloat16, manual cast: torch.bfloat16
    model_type FLUX
    VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
    no CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
    Requested to load LTXAVTEModel_
    FETCH ComfyRegistry Data: 65/135
    loaded partially; 18882.76 MB usable, 18320.26 MB loaded, 7121.04 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
    loaded partially; 18567.87 MB usable, 18005.37 MB loaded, 7435.37 MB offloaded, 562.50 MB buffer reserved, lowvram patches: 0
    Requested to load VideoVAE
    Unloaded partially: 284.24 MB freed, 17721.12 MB remains loaded, 562.50 MB buffer reserved, lowvram patches: 0
    loaded completely; 1616.34 MB usable, 1384.94 MB loaded, full load: True
    FETCH ComfyRegistry Data: 70/135
    Windows fatal exception: access violation

Stack (most recent call first):
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\storage.py", line 466 in getitem
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 136 in load_torch_file
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_lt_audio.py", line 30 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\latest_io.py", line 1764 in EXECUTE_NORMALIZED
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy_api\internal_init_.py", line 149 in wrapped_func
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 296 in process_inputs
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 308 in _async_map_node_over_list
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 334 in get_output_data
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 525 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 758 in execute_async
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\events.py", line 88 in _run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 1986 in _run_once
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 641 in run_forever
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\windows_events.py", line 322 in run_forever
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\base_events.py", line 674 in run_until_complete
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 118 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\asyncio\runners.py", line 194 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\execution.py", line 702 in execute
File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\main.py", line 276 in prompt_worker
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1012 in run
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1075 in _bootstrap_inner
File "H:\ai\comfyui2\ComfyUI_windows_portable\python_embeded\Lib\threading.py", line 1032 in _bootstrap

H:\ai\comfyui2\ComfyUI_windows_portable>pause
请按任意键继续. . .

[rattus128]

That looks like the legacy safetensors fallback. Your startup log indicates pytorch 2.7 which is why aimdo isnt working (and not in play). I have line 136 as non-aimdo code:

File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 136 in load_torch_file

            if comfy.memory_management.aimdo_enabled:
                sd, metadata = load_safetensors(ckpt)
                if not return_metadata:
                    metadata = None
            else:
                with safetensors.safe_open(ckpt, framework="pt", device=device.type) as f:
                    sd = {}
                    for k in f.keys():
(136)                   tensor = f.get_tensor(k) 
                        if DISABLE_MMAP:  # TODO: Not sure if this is the best way to bypass the mmap issues
                            tensor = tensor.to(device=device, copy=True)
                        sd[k] = tensor
                    if return_metadata:
                        metadata = f.metadata()

It doesnt explain why --disable-mmap works for you yet.

Can you try updating your pytorch to a more recent version?

[304523475]

That looks like the legacy safetensors fallback. Your startup log indicates pytorch 2.7 which is why aimdo isnt working (and not in play). I have line 136 as non-aimdo code:看起来像是旧版 safetensors 的回退方案。你的启动日志显示 PyTorch 版本为 2.7，这就是为什么 aimdo 无法工作（并且没有被启用）的原因。我的代码第 136 行是非 aimdo 代码：

File "H:\ai\comfyui2\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 136 in load_torch_file

            if comfy.memory_management.aimdo_enabled:
                sd, metadata = load_safetensors(ckpt)
                if not return_metadata:
                    metadata = None
            else:
                with safetensors.safe_open(ckpt, framework="pt", device=device.type) as f:
                    sd = {}
                    for k in f.keys():
(136)                   tensor = f.get_tensor(k) 
                        if DISABLE_MMAP:  # TODO: Not sure if this is the best way to bypass the mmap issues
                            tensor = tensor.to(device=device, copy=True)
                        sd[k] = tensor
                    if return_metadata:
                        metadata = f.metadata()

It doesnt explain why --disable-mmap works for you yet.它仍然无法解释为什么 --disable-mmap 对你有效。

Can you try updating your pytorch to a more recent version?您能否尝试将 PyTorch 更新到更新的版本？

上面是我在启动bat的命令里没有加--disable-mmap，最开始我有加--disable-mmap，因为后台崩溃所以加了--disable-mmap，但是还是崩溃，所以--disable-mmap不管用。我的touch是2.7.1+cu128，如果更新的话可能会导致我的环境轮子都得更新，这太可怕了。没关系，先将就用吧，谢了