Migrate TBE benchmark utilities over to TBE, pt 5

q10 · facebook-github-bot · commit 05d089a38328 · 2025-03-12T01:16:26.000-07:00
Summary:
- Break up `split_table_batched_embeddings_benchmark.py` into smaller benchmark
  files

Differential Revision: D70924409
diff --git a/fbgemm_gpu/bench/tbe/tbe_inference_benchmark.py b/fbgemm_gpu/bench/tbe/tbe_inference_benchmark.py
@@ -48,6 +48,7 @@
 from torch.profiler import profile
 
 logging.basicConfig(level=logging.DEBUG)
+reporter = BenchmarkReporter(True)
 
 
 def kineto_trace_profiler(p: profile, trace_info: tuple[str, str, str, str]) -> float:
