juliensimon

from keras.models import save_mxnet_model

prefix      = '/opt/ml/'
model_path  = os.path.join(prefix, 'model')

model_name = 'mnist-cnn-'+str(epochs)
model.save(model_path+'/'+model_name+'.hd5') # Keras model
save_mxnet_model(model=model, prefix=model_path+'/'+model_name) # MXNet model

juliensimon

from keras.utils import multi_gpu_model

model = Sequential()
model.add(...)
...
if gpu_count > 1:
    model = multi_gpu_model(model, gpus=gpu_count)
model.compile(...)

juliensimon

def load_data(input_path):
    # Adapted from https://github.com/keras-team/keras/blob/master/keras/datasets/fashion_mnist.py
    files = ['training/train-labels-idx1-ubyte.gz', 
             'training/train-images-idx3-ubyte.gz',
             'validation/t10k-labels-idx1-ubyte.gz', 
             'validation/t10k-images-idx3-ubyte.gz']
    # Load training labels
    with gzip.open(input_path+files[0], 'rb') as lbpath:
        y_train = np.frombuffer(
            lbpath.read(), np.uint8, offset=8)

juliensimon

prefix     = '/opt/ml/'
param_path = os.path.join(prefix, 'input/config/hyperparameters.json')

with open(param_path, 'r') as params:
    hyperParams = json.load(params)
    
lr = float(hyperParams.get('lr', '0.1'))
batch_size = int(hyperParams.get('batch_size', '128'))
epochs = int(hyperParams.get('epochs', '10'))
gpu_count = int(hyperParams.get('gpu_count', '0'))

juliensimon

output_path = 's3://{}/{}/output'.format(sess.default_bucket(), repo_name)
image_name  = '{}.dkr.ecr.{}.amazonaws.com/{}:latest'.format(account, region, repo_name)

estimator = sagemaker.estimator.Estimator(
                       image_name=image_name,
                       base_job_name=base_job_name,
                       role=role, 
                       train_instance_count=1, 
                       train_instance_type=train_instance_type,
                       output_path=output_path,

juliensimon

local_directory = 'data'
prefix          = repo_name+'/input'

train_input_path      = sess.upload_data(
  local_directory+'/train/', key_prefix=prefix+'/train')
validation_input_path = sess.upload_data(
  local_directory+'/validation/', key_prefix=prefix+'/validation')

juliensimon

docker build -t $image_tag -f $dockerfile .
docker tag $image_tag $account.dkr.ecr.$region.amazonaws.com/$repo_name:latest
docker push $account.dkr.ecr.$region.amazonaws.com/$repo_name:latest

juliensimon

aws ecr describe-repositories --repository-names $repo_name > /dev/null 2>&1
if [ $? -ne 0 ]
then
    aws ecr create-repository --repository-name $repo_name > /dev/null
fi

$(aws ecr get-login --region $region --no-include-email)

juliensimon

FROM nvidia/cuda:9.0-runtime

RUN apt-get update && \
    apt-get -y install build-essential libopencv-dev libopenblas-dev libjemalloc-dev libgfortran3 \
    python-dev python3-dev python3-pip wget curl

COPY mnist_cnn.py /opt/program/train
RUN chmod +x /opt/program/train

RUN mkdir /root/.keras

juliensimon

{
    "epsilon": 1e-07, 
    "floatx": "float32", 
    "image_data_format": "channels_first", 
    "backend": "mxnet"
}