Have a benchmark for AutoML

mlcommons / training_policies

Issues related to MLPerf™ training policies, including rules and suggested changes

https://mlcommons.org/en/groups/training

Apache License 2.0

91 stars 65 forks source link

Open ddutta opened 6 years ago

ddutta commented 6 years ago

Have a simple starter benchmark for comparing different AutoML algorithms e.g. from http://www.ml4aad.org/literature-on-neural-architecture-search/

Supporting orgs: Cisco

petermattson commented 6 years ago

Hi Debo, this would be separate from MLPerf training benchmark suite correct? Just supported by MLPerf org?

ddutta commented 5 years ago

Yes ... it would cover things like: ENAS, MnasNet, AmoebaNet, DARTS etc.