So like the subject suggests, I accidentally installed some 3000+ custom nodes, or at least started to. I hit stop when I noticed what was happening, but not before installing hundreds. the first time I booted comfyui it ran scripts for like 45 mins, but now it just keeps saying “unable to start ComfyUI” I’ll post the log below. What can I do?
[2025-07-17 22:11:12.059] [info]
Prestartup times for custom nodes:
3.3 seconds: C:\Users\pyrob\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\custom_nodes\ComfyUI-Manager
[2025-07-17 22:11:12.059] [error] expected str, bytes or os.PathLike object, not NoneType
[2025-07-17 22:11:13.198] [info] Checkpoint files will always be loaded safely.
[2025-07-17 22:11:13.208] [error] Traceback (most recent call last):
[2025-07-17 22:11:13.210] [error] File "C:\Users\pyrob\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\main.py", line 132, in <module>
import execution
File "C:\Users\pyrob\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\execution.py", line 14, in <module>
import comfy.model_management
File "C:\Users\pyrob\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\comfy\model_management.py", line 221, in <module>
total_vram = get_total_memory(get_torch_device()) / (1024 * 1024)
[2025-07-17 22:11:13.210] [error] ^^^^^^^^^^^^^^^^^^
File "C:\Users\pyrob\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\comfy\model_management.py", line 172, in get_torch_device
[2025-07-17 22:11:13.211] [error] return torch.device(torch.cuda.current_device())
[2025-07-17 22:11:13.212] [error] ^^^^^^^^^^^^^^^
[2025-07-17 22:11:13.212] [error] ^^^^^^^^^^^^
[2025-07-17 22:11:13.213] [error] File "C:\Users\pyrob\Documents\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 1026, in current_device
_lazy_init()
File "C:\Users\pyrob\Documents\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 363, in _lazy_init
[2025-07-17 22:11:13.214] [error] raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled