Failed to install opencv in a step of a pipeline

I am trying to make a trining step for a pipeline where I need open cv, I have manage to install it in the environment but for some reason it doesn't find it.
My error:

Execution failed. User process '/azureml-envs/azureml***/bin/python' exited with status code 1. Please check log file 'user_logs/std_log.txt' for error details. Error: Traceback (most recent call last):
File "train.py", line 8, in
import cv2
File "/azureml-envs/azureml***/lib/python3.8/site-packages/cv2/__init__.py", line 181, in
bootstrap()
File "/azureml-envs/azureml***/lib/python3.8/site-packages/cv2/__init__.py", line 153, in bootstrap
native_module = importlib.import_module("cv2")
File "/azureml-envs/azureml***/lib/python3.8/importlib/__init__.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
ImportError: libGL.so.1: cannot open shared object file: No such file or directory

# I have also tried with regular import opencv-python
my yml:
name: train_env
channels:
- conda-forge
dependencies:
- python=3.8.1
- ipykernel
- PyYAML
- requests
- tqdm
- pip
- pip:
- opencv-python-headless
- azureml-defaults
- torch
- torchvision
- ultralytics
- matplotlib
- pandas
- pyarrow
- Pillow
- onnx
- pathlib
- numpy
- requests
- scipy

提供された environment yml と train.py から始め、次に cv2 をインポートする Azure Machine Learning パイプラインのステップを再現します。libGL.so.1 が見つからないエラーの前後にある環境ログを確認します。パイプライン環境で cv2 をインポートでき、そのトレーニングステップを通過して続行できれば完了です。

索引モデルが issue の本文から書いたものです。