For training with PyTorch, please visit PyTorch Encoding Toolkit
# assuming you have CUDA 10.0 on your machine
pip install mxnet-cu100
HOROVOD_GPU_ALLREDUCE=NCCL pip install -v --no-cache-dir horovod
pip install --no-cache mpi4py
-
Unfortunately ,this is required for training using MXNet Gluon. Please follow the GluonCV tutorial to prepare the data.
-
Copy the data into ramdisk (optional):
cd ~/ sudo mkdir -p /media/ramdisk sudo mount -t tmpfs -o size=200G tmpfs /media/ramdisk cp -r /home/ubuntu/data/ILSVRC2012/ /media/ramdisk
Using ResNeSt-50 as the target model:
horovodrun -np 64 --hostfile hosts python train.py \
--rec-train /media/ramdisk/ILSVRC2012/train.rec \
--rec-val /media/ramdisk/ILSVRC2012/val.rec \
--model resnest50 --lr 0.05 --num-epochs 270 --batch-size 128 \
--use-rec --dtype float32 --warmup-epochs 5 --last-gamma --no-wd \
--label-smoothing --mixup --save-dir params_ resnest50 \
--log-interval 50 --eval-frequency 5 --auto_aug --input-size 224
python verify.py --model resnest50 --crop-size 224 --resume params_ resnest50/imagenet-resnest50-269.params