We need to collect, prepare, and present our results so far for the Milestone 2.
Broadly, I believe we need to:
Compare DARTS architecture learned on Graphene (#6) to hand-designed architectures (#3), and compare/contrast performance (accuracy, training time, number of parameters, etc)
Compare DARTS performance/architecture on Graphene to random search (#5)
We need to collect, prepare, and present our results so far for the Milestone 2.
Broadly, I believe we need to:
Compare DARTS architecture learned on Graphene (#6) to hand-designed architectures (#3), and compare/contrast performance (accuracy, training time, number of parameters, etc)
Compare DARTS performance/architecture on Graphene to random search (#5)