Slow Startup When Using VSCode Debugger with Matplotlib Installation via CLI Model Configuration in PyTorch Lightning

[lengmo1996]

Bug description

When configuring models using the command-line interface (CLI) in PyTorch Lightning and attempting to debug with VSCode Debugger, the startup process is significantly slow. This issue specifically occurs during the initialization phase of the debugging session.

What version are you seeing the problem on?

v2.5

How to reproduce the bug

1. Pull a docker image from docker hub by 'docker pull pytorchlightning/pytorch_lightning:2.5.1-py3.12-torch2.5-cuda12.1.1'
2. Create a container from the image.
3. Connect the container by vscode and install some python extensions.
4. Set up vscode debugger for debugging
5. Initiate a debug sesscion to debug the following python code.

# from lightning.pytorch.demos import BoringDataModule, BoringModel
from lightning.pytorch.cli import (
    LightningCLI,
)

if __name__ == "__main__":
    LightningCLI()

Error messages and logs

# Error messages and logs here please

Environment

Current environment

• VSCODE：
  - Version: 1.97.2 (system setup)
  - Submit: e54c774e0add60467559eb0d1e229c6452cf8447
  - Date: 2025-02-12T23:20:35.343Z
  - Electron: 32.2.7
  - ElectronBuildId: 10982180
  - Chromium: 128.0.6613.186
  - Node.js: 20.18.1
  - V8: 12.8.374.38-electron.0
  - OS: Windows_NT x64 10.0.19045
• VSCODE Extension：
  - Pylance (ms-python.vscode-pylance): 2025.4.1
  - Python (ms-python.python): 2024.4.1
  - Python Debugger (ms-python.debugpy): 2025.6.0
• CUDA:
  - GPU:
  - NVIDIA GeForce RTX 3090
  - NVIDIA GeForce RTX 3090
  - available: True
  - version: 12.1
• Lightning:
  - lightning: 2.5.1
  - lightning-utilities: 0.11.9
  - pytorch-lightning: 2.5.1
  - torch: 2.5.1+cu121
  - torchmetrics: 1.4.3
• Packages:
  - absl-py: 2.1.0
  - aiohappyeyeballs: 2.6.1
  - aiohttp: 3.11.14
  - aiosignal: 1.3.2
  - antlr4-python3-runtime: 4.9.3
  - anyio: 4.9.0
  - attrs: 25.3.0
  - autocommand: 2.2.2
  - backports.tarfile: 1.2.0
  - bitsandbytes: 0.45.2
  - blinker: 1.4
  - cachetools: 5.5.2
  - certifi: 2025.1.31
  - charset-normalizer: 3.4.1
  - click: 8.1.8
  - cloudpickle: 2.2.1
  - coloredlogs: 15.0.1
  - contourpy: 1.3.1
  - coverage: 7.3.1
  - cryptography: 3.4.8
  - cycler: 0.12.1
  - dbus-python: 1.2.18
  - deepspeed: 0.9.3
  - distro: 1.7.0
  - docstring-parser: 0.16
  - fastapi: 0.115.11
  - filelock: 3.18.0
  - flatbuffers: 25.2.10
  - fonttools: 4.56.0
  - frozenlist: 1.5.0
  - fsspec: 2024.3.1
  - google-auth: 2.38.0
  - google-auth-oauthlib: 1.0.0
  - grpcio: 1.71.0
  - h11: 0.14.0
  - hjson: 3.1.0
  - httplib2: 0.20.2
  - humanfriendly: 10.0
  - hydra-core: 1.3.2
  - idna: 3.10
  - importlib-metadata: 4.6.4
  - importlib-resources: 6.5.2
  - inflect: 7.3.1
  - iniconfig: 2.0.0
  - jaraco.collections: 5.1.0
  - jaraco.context: 5.3.0
  - jaraco.functools: 4.0.1
  - jaraco.text: 3.12.1
  - jeepney: 0.7.1
  - jinja2: 3.1.6
  - joblib: 1.4.2
  - jsonargparse: 4.35.0
  - keyring: 23.5.0
  - kiwisolver: 1.4.8
  - launchpadlib: 1.10.16
  - lazr.restfulclient: 0.14.4
  - lazr.uri: 1.0.6
  - lightning: 2.5.1
  - lightning-utilities: 0.11.9
  - markdown: 3.7
  - markdown-it-py: 3.0.0
  - markupsafe: 3.0.2
  - matplotlib: 3.10.1
  - mdurl: 0.1.2
  - more-itertools: 8.10.0
  - mpmath: 1.3.0
  - multidict: 6.2.0
  - networkx: 3.4.2
  - ninja: 1.11.1.3
  - numpy: 1.26.4
  - nvidia-cublas-cu12: 12.1.3.1
  - nvidia-cuda-cupti-cu12: 12.1.105
  - nvidia-cuda-nvrtc-cu12: 12.1.105
  - nvidia-cuda-runtime-cu12: 12.1.105
  - nvidia-cudnn-cu12: 9.1.0.70
  - nvidia-cufft-cu12: 11.0.2.54
  - nvidia-curand-cu12: 10.3.2.106
  - nvidia-cusolver-cu12: 11.4.5.107
  - nvidia-cusparse-cu12: 12.1.0.106
  - nvidia-nccl-cu12: 2.21.5
  - nvidia-nvjitlink-cu12: 12.8.93
  - nvidia-nvtx-cu12: 12.1.105
  - oauthlib: 3.2.0
  - omegaconf: 2.3.0
  - onnx: 1.16.2
  - onnxruntime: 1.18.1
  - packaging: 23.1
  - pandas: 2.1.4
  - pillow: 11.1.0
  - pip: 25.0.1
  - platformdirs: 4.2.2
  - pluggy: 1.5.0
  - propcache: 0.3.0
  - protobuf: 6.30.1
  - psutil: 5.9.5
  - py-cpuinfo: 9.0.0
  - pyasn1: 0.6.1
  - pyasn1-modules: 0.4.1
  - pydantic: 1.10.21
  - pygments: 2.19.1
  - pygobject: 3.42.1
  - pyjwt: 2.3.0
  - pyparsing: 2.4.7
  - pytest: 7.4.0
  - pytest-cov: 4.1.0
  - pytest-random-order: 1.1.0
  - pytest-rerunfailures: 12.0
  - pytest-timeout: 2.1.0
  - python-apt: 2.4.0+ubuntu4
  - python-dateutil: 2.9.0.post0
  - pytorch-lightning: 2.5.1
  - pytz: 2025.1
  - pyyaml: 6.0.2
  - requests: 2.32.3
  - requests-oauthlib: 2.0.0
  - rich: 13.5.3
  - rsa: 4.9
  - scikit-learn: 1.3.2
  - scipy: 1.15.2
  - secretstorage: 3.3.1
  - setuptools: 75.6.0
  - six: 1.16.0
  - sniffio: 1.3.1
  - starlette: 0.46.1
  - sympy: 1.13.1
  - tensorboard: 2.14.1
  - tensorboard-data-server: 0.7.2
  - tensorboardx: 2.6.2.2
  - threadpoolctl: 3.6.0
  - tomli: 2.0.1
  - torch: 2.5.1+cu121
  - torchmetrics: 1.4.3
  - tqdm: 4.66.6
  - triton: 3.1.0
  - typeguard: 4.3.0
  - typeshed-client: 2.7.0
  - typing-extensions: 4.10.0
  - tzdata: 2025.1
  - urllib3: 2.3.0
  - uvicorn: 0.34.0
  - wadllib: 1.3.6
  - werkzeug: 3.1.3
  - wget: 3.2
  - wheel: 0.43.0
  - yarl: 1.18.3
  - zipp: 1.0.0
• System:
  - OS: Linux
  - architecture:
  - 64bit
  - ELF
  - processor: x86_64
  - python: 3.12.9
  - release: 5.15.0-131-generic
  - version: Error if dataset size = 1 batch. #141~20.04.1-Ubuntu SMP Thu Jan 16 18:38:51 UTC 2025

More info

No response

cc @mauvilsa

[lengmo1996]

I'm sorry but I forgot to mention that this issue is closely related to the matplotlib package. If matplotlib is removed from the environment, this problem won't occur. However, I don't understand the underlying mechanism behind this.

[SkafteNicki]

I'm sorry but I forgot to mention that this issue is closely related to the matplotlib package. If matplotlib is removed from the environment, this problem won't occur. However, I don't understand the underlying mechanism behind this.

If removing matplotlib from the environment solves the problem, I highly doubt that this is actually a problem with lightning and more between vscode debugger and matplotlib. Can you please confirm that if you have a environment without lightning but still having matplotlib that the issue still persist or not?

[lengmo1996]

Thanks a lot for your reply and clarification!
I also tested with only matplotlib (without lightning) and could not reproduce the slowdown, so the issue seems to occur only when both lightning and matplotlib are present.
However, due to recent work commitments, I don’t have time to perform a stable reproduction at the moment. Please feel free to close this issue for now, and I will reopen it if I’m able to reproduce the problem consistently in the future.
Really appreciate your time and support!

[SkafteNicki]

@lengmo1996 yeah sorry for first replying to your message after a couple of months.
Alright, thanks for confirming that it only happens when lightning and matplotlib is present in a environment.
I am going to close the issue, mostly because it seems to be an edgecase involving multiple packages and vscode debugger. No matter what the underlying reason is, it will be low priority to debug.
But yes feel free to ping me and reopen if necessary.