- Use architecture of networks published in the literature
- Use open source implementation if possible
- Pre-trained models and fine turn your dataset


University of Michigan - Ann Arbor

Benchmarking is a way to analyze how well your neural network is performing on a given standard set.
-Ensembling
-Multicrop at test time
-using open source code


Benchmarking

Trains several networks independently and average their outputs


Ensembling

Run classifier on multiple versions of test images and average results


Learn Before

Related