search

mlcommons / training

Reference implementations of MLPerf™ training benchmarks
https://mlcommons.org/en/groups/training
Apache License 2.0
1.57k stars 548 forks source link

Pack image classification benchmark on MLCube #508

Open davidjurado opened 2 years ago

davidjurado commented 2 years ago

Benchmark execution with MLCube

Project setup

# Create Python environment and install MLCube Docker runner 
virtualenv -p python3 ./env && source ./env/bin/activate && pip install mlcube-docker

# Fetch the RNN speech recognition workload
git clone https://github.com/mlcommons/training && cd ./training
git fetch origin pull/508/head:feature/mlcube_image_classification
git checkout feature/mlcube_image_classification && cd ./image_classification/mlcube

Dataset

The ImageNet needs to be downloaded manually.

Tasks execution

# Download ImageNet dataset. Default path = /workspace/data
# To override it, use data_dir=DATA_DIR
mlcube run --task download

By default MLCube images use pull-type installation, so they should be available on docker hub. If not, try this:

mlcube run ... -Pdocker.build_strategy=auto

We are targeting pull-type installation, so MLCube images should be available on docker hub. If not, try this:

mlcube run ... -Pdocker.build_strategy=always
github-actions[bot] commented 2 years ago

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

johntran-nv commented 1 year ago

@sgpyc could you please review when you get a chance?