- Install Torch7
- Install CUDA libraries and make sure you can used cudnn
- Install the 3rd Party Scripts
./install_thirdparty.sh
The third party scripts are installed as part of the jeepers common workflow scripts.
Added from monitoring tools from trainplot
There is also a transform.lua from facebook that does the data augmentation
As jeepers and the install process improves, this should be less confusing.
There is some setup required to prepare your data. Below we assume you have a directory above this repository called "kaggle_data", where images are already converted to png and there is a csv with labels.
CUDA_VISIBLE_DEVICES=6,7 th train.lua -lr 0.0005 -plotid 1 -out_path_base output/checkpoints/small_training/small_multigpu_test -val 1 -train_dir ../kaggle_data/train_small_90val_png/ -test_dir ../kaggle_data/train_small_10val_png/ -input_size 128 -batch_size 256 -resample_weight 0.9 -pool_size 12 -multi_gpu 1
This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.