Making the code dataset agnostic

quark0 / darts

Differentiable architecture search for convolutional and recurrent networks

https://arxiv.org/abs/1806.09055

Apache License 2.0

3.92k stars 843 forks source link

Making the code dataset agnostic #119

Closed dylanrandle closed 5 years ago

dylanrandle commented 5 years ago

This PR adds a datasets module with a load_dataset function that loads datasets, preprocesses them, and returns various important parameters (e.g. num channels, output dimension) to perform training.

This PR also makes the model_search.Network and model.NetworkCifar objects agnostic to the number of input channels of the data (as far as I'm aware).

dylanrandle commented 5 years ago

Wrong place