6124-add-training-attribute-check

[binliunls]

Fixes #6124 .

Description

When running the inference with torchscript wrapped TensorRT models, the evaluator would give an error. This is caused by the with engine.mode() code run the training method of engine.network without checking. In this PR, an attribute check has been added to cover this issue.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[wyli]

I think the root cause is here n.training is used without checking:

MONAI/monai/networks/utils.py

Line 379 in 4170924

training = [n for n in nets if n.training]

perhaps we fix that utility function? also this

MONAI/monai/networks/utils.py

Line 413 in 4170924

eval_list = [n for n in nets if not n.training]

[wyli]

/build