Correctly handle OSError in test_bench script

[xuzhao9]

Handle subprocess segmentation fault error in the test_bench script.

Example:

python run_benchmark.py test_bench -m yolov3 -d cuda -t train

{
    "name": "test_bench",
    "environ": {
        "pytorch_git_version": "f13a21cce50586b74c882ec8e9ef38d60458484e",
        "pytorch_version": "2.2.0.dev20231213+cu118",
        "device": "Tesla T4"
    },
    "metrics": {
        "model=yolov3, test=eval, device=cuda, bs=None, extra_args=[], metric=latencies": "Subprocess terminates with code 11",
        "model=yolov3, test=eval, device=cuda, bs=None, extra_args=[], metric=cpu_peak_mem": "Subprocess terminates with code 11",
        "model=yolov3, test=eval, device=cuda, bs=None, extra_args=[], metric=gpu_peak_mem": "Subprocess terminates with code 11"
    }
}

Nightly docker build workflow: https://github.com/pytorch/benchmark/actions/runs/7257082609

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[aaronenyeshi]

LGTM!

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@xuzhao9 merged this pull request in c64a436.