/imagenet_training

Pytorch code for training imagenet with fp16

Primary LanguagePythonMIT LicenseMIT

ImageNet Training

Pytorch code for training imagenet with fp16

  1. Install pytorch,torchvision
  2. (Optional) install data loading speedups:
conda install -c thomasbrandon -c defaults -c conda-forge pillow-accel-avx2
conda install -c conda-forge libjpeg-turbo

usage example:

CUDA_VISIBLE_DEVICES=1,0 python3 train_imagenet.py -t /imagenet_data/train/ -v  /imagenet_data/val/ -a resnet152 -b 256 -j=6 -n=my_experiment_name --lr=.1 -fp16