kelvingakuo/Agrix-app-demo

SETTIING UP THE ENVIRONMENT

1. Create a virtualenv e.g. using Conda, running Python 3.6

conda create -n nameofEnv python=3.6

2. Activate the env

source activate nameofEnv

3. Install requirements

pip install -r requirements.txt

4. Create needed dirs

mkdir data/crowdai_train
mkdir data/crowdai_test
mkdir saved_models

DATA

1. Download train.tar and test.tar from crowdai.org/challenges/1 under 'Datasets'
2. Extract train.tar into 'data/crowdai_train' and test.tar into 'data/crowdai_test'

USAGE

1. Run

python train_nn.py modelName

The modelName is either 'alexnet' or 'vgg16'

2. Wait