Skip to content

generate-nvidia-engine

Automatically generated README for this automation recipe: generate-nvidia-engine

Category: MLPerf benchmark support

License: Apache 2.0

This CM script is in draft stage

  • CM meta description for this script: _cm.yaml
  • Output cached? False

Reuse this script in your project

Install MLCommons CM automation meta-framework

Pull CM repository with this automation recipe (CM script)

cm pull repo mlcommons@cm4mlops

cmr "generate engine mlperf inference nvidia" --help

Run this script

Run this script via CLI
cm run script --tags=generate,engine,mlperf,inference,nvidia[,variations] [--input_flags]
Run this script via CLI (alternative)
cmr "generate engine mlperf inference nvidia [variations]" [--input_flags]
Run this script from Python
import cmind

r = cmind.access({'action':'run'
              'automation':'script',
              'tags':'generate,engine,mlperf,inference,nvidia'
              'out':'con',
              ...
              (other input keys for this script)
              ...
             })

if r['return']>0:
    print (r['error'])
Run this script via Docker (beta)
cm docker script "generate engine mlperf inference nvidia[variations]" [--input_flags]

Variations

  • No group (any combination of variations can be selected)

    Click here to expand this section.

    • _batch_size.#
      • ENV variables:
        • CM_MODEL_BATCH_SIZE: None
    • _copy_streams.#
      • ENV variables:
        • CM_GPU_COPY_STREAMS: None
    • _cuda
      • ENV variables:
        • CM_MLPERF_DEVICE: gpu
        • CM_MLPERF_DEVICE_LIB_NAMESPEC: cudart

  • Group "device"

    Click here to expand this section.

    • _cpu (default)
      • ENV variables:
        • CM_MLPERF_DEVICE: cpu

  • Group "model"

    Click here to expand this section.

    • _resnet50 (default)
      • ENV variables:
        • CM_MODEL: resnet50
    • _retinanet
      • ENV variables:
        • CM_MODEL: retinanet
Default variations

_cpu,_resnet50

Script flags mapped to environment

  • --output_dir=value → CM_MLPERF_OUTPUT_DIR=value

Default environment

These keys can be updated via --env.KEY=VALUE or env dictionary in @input.json or using script flags.

  • CM_BATCH_COUNT: 1
  • CM_BATCH_SIZE: 1
  • CM_LOADGEN_SCENARIO: Offline
  • CM_GPU_COPY_STREAMS: 1
  • CM_TENSORRT_WORKSPACE_SIZE: 4194304

Native script being run

No run file exists for Windows

Script output

cmr "generate engine mlperf inference nvidia [variations]" [--input_flags] -j