use transformers for now

optimum_benchmark/cli.py

@@ -53,23 +53,14 @@
 cs.store(group="launcher", name=ProcessConfig.name, node=ProcessConfig)
 cs.store(group="launcher", name=TorchrunConfig.name, node=TorchrunConfig)
 
-LOGGING_SETUP_DONE = False
-
-
-def setup_logging_once(*args, **kwargs):
-    global LOGGING_SETUP_DONE
-    if not LOGGING_SETUP_DONE:
-        LOGGING_SETUP_DONE = True
-        setup_logging(*args, **kwargs)
-
 
 # optimum-benchmark
 @hydra.main(version_base=None)
 def main(config: DictConfig) -> None:
     log_level = os.environ.get("LOG_LEVEL", "INFO")
     log_to_file = os.environ.get("LOG_TO_FILE", "1") == "1"
     override_benchmarks = os.environ.get("OVERRIDE_BENCHMARKS", "0") == "1"
-    setup_logging_once(level=log_level, to_file=log_to_file, prefix="MAIN-PROCESS")
+    setup_logging(level=log_level, to_file=log_to_file, prefix="MAIN-PROCESS")
 
     if glob.glob("benchmark_report.json") and not override_benchmarks:
         LOGGER.warning(

optimum_benchmark/task_utils.py

@@ -47,17 +47,12 @@
 _TIMM_TASKS_TO_MODEL_LOADERS = {
     "image-classification": "create_model",
 }
-_SENTENCE_TRANSFORMERS_TASKS_TO_MODEL_LOADERS = {
-    "feature-extraction": "SentenceTransformer",
-    "sentence-similarity": "SentenceTransformer",
-}
 
 
 _LIBRARY_TO_TASKS_TO_MODEL_LOADER_MAP = {
     "timm": _TIMM_TASKS_TO_MODEL_LOADERS,
     "diffusers": _DIFFUSERS_TASKS_TO_MODEL_LOADERS,
     "transformers": _TRANSFORMERS_TASKS_TO_MODEL_LOADERS,
-    "sentence-transformers": _SENTENCE_TRANSFORMERS_TASKS_TO_MODEL_LOADERS,
 }
 
 _SYNONYM_TASK_MAP = {
@@ -122,6 +117,9 @@ def infer_library_from_model_name_or_path(model_name_or_path: str, revision: Opt
     if inferred_library_name is None:
         raise KeyError(f"Could not find the proper library name for {model_name_or_path}.")
 
+    if inferred_library_name == "sentence-transformers":
+        inferred_library_name = "transformers"
+
     return inferred_library_name