Accelerated by another 5% using FP16/32 Batch-norm for Tensor Cores.

AlexeyAB

2018-04-16 9bae70b22549b68f5cdeece8b6c3b3de00c22714

1
2
3
4
5
6
7
classes=1000
train  = /data/imagenet/imagenet1k.train.list
valid  = /data/imagenet/imagenet1k.valid.list
backup = /home/pjreddie/backup/
labels = data/imagenet.labels.list
names  = data/imagenet.shortnames.list
top=5