Jean-Tsung Hsiao

While running "sh ./reference_mlperf_accuracy.sh" got this warning: WARNING 09-15 23:53:29 [arg_utils.py:1770] ['WhisperForConditionalGeneration'] is not supported by the V1 Engine. Falling back to V0.

Initial part of mlperf_log_detail.txt: :::MLLOG {"key": "loadgen_version", "value": "5.1.0 @ 50de99161e", "time_ms": 0.006612, "namespace": "mlperf::logging", "event_type": "POINT_IN_TIME", "metadata": {"is_error": false, "is_warning": false, "file": "version.cc", "line_no": 53, "pid": 147525, "tid": 147525}}...

Initial part of the console log: (mlperf) [email protected]:/work/build/inference/speech2text$ head -f -n 500 mlper_loadgen_response.log head: invalid option -- 'f' Try 'head --help' for more information. (mlperf) [email protected]:/work/build/inference/speech2text$ head -h head: invalid...

Hi Arjun, Yes, that fixed the issue. Thanks! Two questions about running "sh ./reference_mlperf_accuracy.sh". The test bed has two A100 gpus. First of all, most of the time only gpu-0...