args = parser.parse_args()
if not os.path.exists(args.output):
os.makedirs(args.output)
print('-------------')
print('BATCH : {}'.format(args.batch))
print('EPOCH : {}'.format(args.epoch))
print('ALPA : {}'.format(args.alpha))
print('DROPOUT : {}'.format(args.dropout))
print('Load Weights?: {}'.format(args.weights))
print('Dataset : {}'.format(args.dataset))
print('OUTPUT : {}'.format(args.output))
print('-------------')
df_train, df_val = getDataFromFolder(args.dataset, args.output)
print('TRAIN:', len(df_train))
print('VALIDATION:', len(df_val))
model = NvidiaModel(args.alpha, args.dropout)
print(model.summary())
# Saves the model...
with open(os.path.join(args.output, 'model.json'), 'w') as f:
f.write(model.to_json())
try:
if args.weights:
print('Loading weights from file ...')
model.load_weights(args.weights)
print('BATCH: {}'.format(args.batch))
print('EPOCH: {}'.format(args.epoch))
print('ALPA: {}'.format(args.alpha))
print('DROPOUT: {}'.format(args.dropout))
print('Load Weights?: {}'.format(args.loadWeights))
print('Dataset: {}'.format(args.dataset))
print('Model: {}'.format(args.output))
print('-------------')
if not os.path.exists(args.output):
os.makedirs(args.output)
# ROOT = '/Users/nando/Downloads/thunderhill_data/dataset_sim_000_km_few_laps'
# split data into training and testing
# df_train, df_val = __train_test_split('{}/driving_log.csv'.format(ROOT), False)
df_train, df_val = getDataFromFolder(args.dataset)
print('TRAIN:', len(df_train))
print('VALIDATION:', len(df_val))
model = NvidiaModel(args.alpha, args.dropout)
# Saves the model...
with open(os.path.join(args.output, 'model.json'), 'w') as f:
f.write(model.to_json())
try:
if args.weights:
print('Loading weights from file ...')
model.load_weights(args.weights)
except IOError:
print("No model found")
if not os.path.exists(args.output):
os.makedirs(args.output)
print('-------------')
print('BATCH : {}'.format(args.batch))
print('EPOCH : {}'.format(args.epoch))
print('ALPA : {}'.format(args.alpha))
print('DROPOUT : {}'.format(args.dropout))
print('Load Weights?: {}'.format(args.weights))
print('Dataset : {}'.format(args.dataset))
print('OUTPUT : {}'.format(args.output))
print('-------------')
df = getDataFromFolder(args.dataset,
args.output,
randomize=False,
balance=True,
split=False)
extractCNNFeatures(df, args.weights)
train_size = int(0.7 * len(df))
df_train = df.iloc[:train_size]
df_val = df.iloc[train_size:]
print('TRAIN:', len(df_train))
print('VALIDATION:', len(df_val))
model = Lstm(args.batch, args.seq, CNN_INPUT_SIZE, args.dropout)
print(model.summary())
# Saves the model...
os.makedirs(args.output)
print('-------------')
print('BATCH : {}'.format(args.batch))
print('EPOCH : {}'.format(args.epoch))
print('ALPA : {}'.format(args.alpha))
print('DROPOUT : {}'.format(args.dropout))
print('Load Weights?: {}'.format(args.weights))
print('Dataset : {}'.format(args.dataset))
print('OUTPUT : {}'.format(args.output))
print('-------------')
# TODO: abstract method to normalize speed.
df = getDataFromFolder(args.dataset,
args.output,
randomize=False,
split=True,
normalize=True)[0]
df_train, df_val = getDataFromFolder(args.dataset, args.output)
print('TRAIN:', len(df_train))
print('VALIDATION:', len(df_val))
model = NvidiaModel(args.dropout)
print(model.summary())
# Saves the model...
with open(os.path.join(args.output, 'model.json'), 'w') as f:
f.write(model.to_json())
try:
if args.weights: