[BUG] Tabular models fail to function properly in certain Docker container environments.

Bug Report Checklist

• I provided code that demonstrates a minimal reproducible example.
• [√] I confirmed bug exists on the latest mainline of AutoGluon via source install.
• [√] I confirmed bug exists on the latest stable version of AutoGluon.

Describe the bug

When running tabular models such as DirectTabular or RecursiveTabular in a Docker container where the number of logical CPUs is less than the number of physical CPUs, a bug occurs.

It appears that every tabular model class has overridden a method named _get_default_resources(). For instance, in line 518 of tabular/models/lgb/lgb_model.py, there is an annotation stating that logical=False is faster for training. However, in practice, setting logical=False can lead to bugs in certain Docker environments.

Given this issue, why not change the setting to logical=True, or simply remove this overridden method altogether, since the parent class already implements this method? This adjustment could help avoid the bug and ensure more consistent behavior across different environments.

Expected behavior

To Reproduce

Screenshots / Logs

Installed Versions

# Replace this code with the output of the following:
from autogluon.core.utils import show_versions
show_versions()