Add `--skip` arg for `test_bench` (#2551)
Summary:
### Usage
```
python run_benchmark.py test_bench --accuracy -d mps --models BERT_pytorch,hf_GPT2 --skip BERT_pytorch --output result.json
```
### Result
```
$ python run_benchmark.py test_bench --accuracy -d mps --models BERT_pytorch,hf_GPT2 --skip BERT_pytorch --output result.json
Running TorchBenchModelConfig(name='BERT_pytorch', test='eval', device='mps', batch_size=None, extra_args=['--accuracy'], extra_env=None, output_dir=None, skip=True) ... [skip]
Running TorchBenchModelConfig(name='hf_GPT2', test='eval', device='mps', batch_size=None, extra_args=['--accuracy'], extra_env=None, output_dir=None, skip=False) ... [done]
{
"name": "test_bench",
"environ": {
"pytorch_git_version": "dd2e6d61409aac22198ec771560a38adb0018ba2",
"pytorch_version": "2.6.0.dev20241120"
},
"metrics": {
"model=BERT_pytorch, test=eval, device=mps, bs=None, extra_args=['--accuracy'], metric=accuracy": "skip",
"model=hf_GPT2, test=eval, device=mps, bs=None, extra_args=['--accuracy'], metric=accuracy": "pass"
}
}
```
Pull Request resolved: https://github.com/pytorch/benchmark/pull/2551
Reviewed By: xuzhao9
Differential Revision: D66659994
Pulled By: FindHao
fbshipit-source-id: 7623ba653d7e94866d576d9774ca144bcc657800
