minor fix to filter (#30200)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/30200
as title
Test Plan:
```
buck run mode/opt //caffe2/benchmarks/operator_benchmark:benchmark_all_other_test -- --tag_filter all --iterations 1 --ai_pep_format True --operators None --iterations -1 --warmup_iterations -1 --wipe_cache --forward_only False --device cpu --tag_filter all --use_jit False --operator_range b-z
# ----------------------------------------
# PyTorch/Caffe2 Operator Micro-benchmarks
# ----------------------------------------
# Tag : all
# Benchmarking PyTorch: batchnorm
PyTorchObserver {"type": "PyTorch_batchnorm_M1_N256_K3136_cpu_Eager", "metric": "latency", "unit": "ms", "value": "0.29026457108557224"}
PyTorchObserver {"type": "PyTorch_batchnorm_M1_N256_K3136_cpu_Eager", "metric": "latency", "unit": "ms", "value": "0.2813781425356865"}
PyTorchObserver {"type": "PyTorch_batchnorm_M1_N256_K3136_cpu_Eager", "metric": "latency", "unit": "ms", "value": "0.28009670320898294"}
...
Reviewed By: hl475
Differential Revision: D18627512
fbshipit-source-id: 23f622b96168f90a8d8648bfd9ff9a5116baafdf