From 2ec5974099d3edac5f883b85ff77ad6feab933d2 Mon Sep 17 00:00:00 2001 From: Arjun Suresh Date: Thu, 15 Feb 2024 08:55:37 +0000 Subject: [PATCH] Update README_nvidia.md --- docs/mlperf/inference/gpt-j/README_nvidia.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/docs/mlperf/inference/gpt-j/README_nvidia.md b/docs/mlperf/inference/gpt-j/README_nvidia.md index 7df6e5fb84..0dcbb53ad4 100644 --- a/docs/mlperf/inference/gpt-j/README_nvidia.md +++ b/docs/mlperf/inference/gpt-j/README_nvidia.md @@ -17,6 +17,8 @@ cmr "generate-run-cmds inference _find-performance _all-scenarios" \ ``` * Use `--division=closed` to run all scenarios for the closed division. * Use `--category=datacenter` to run datacenter scenarios +* Use `--model=gptj-99.9` to run the high-accuracy model +* Use `--rerun` to force a rerun even when result files (from a previous run) exist ### Do full accuracy and performance runs for all the scenarios @@ -30,6 +32,8 @@ cmr "generate-run-cmds inference _submission _all-scenarios" --model=gptj-99 \ * Use `--power=yes` for measuring power. It is ignored for accuracy and compliance runs * Use `--division=closed` to run all scenarios for the closed division. No compliance runs are there for gptj. * `--offline_target_qps`, `--server_target_qps`, and `--singlestream_target_latency` can be used to override the determined performance numbers +* Use `--model=gptj-99.9` to run the high-accuracy model +* Use `--rerun` to force a rerun even when result files (from a previous run) exist ### Generate and upload MLPerf submission