flatten = mx.symbol.Flatten(data=avg, name='flatten')
fc1 = mx.symbol.FullyConnected(data=flatten,
num_hidden=nhidden,
name='fc1')
softmax = mx.symbol.Softmax(data=fc1, name='softmax')
return softmax
softmax = inception(21841, 1.0)
batch_size = 64
num_gpu = 4
gpus = [mx.gpu(i) for i in range(num_gpu)]
input_shape = (3, 224, 224)
train = ilsvrc12_iterator(batch_size=batch_size, input_shape=(3, 224, 224))
model_prefix = "model/Inception-Full"
num_round = 10
logging.info(
"This script is used to train ImageNet fullset over 21841 classes.")
logging.info("For noraml 1000 classes problem, please use inception.py")
model = mx.model.FeedForward(ctx=gpus,
symbol=softmax,
num_round=num_round,
learning_rate=0.05,
momentum=0.9,
wd=0.00001)
# global avg pooling
avg = mx.symbol.Pooling(data=in5b, kernel=(7, 7), stride=(1, 1), name="global_pool", pool_type='avg')
# linear classifier
flatten = mx.symbol.Flatten(data=avg, name='flatten')
fc1 = mx.symbol.FullyConnected(data=flatten, num_hidden=nhidden, name='fc1')
softmax = mx.symbol.Softmax(data=fc1, name='softmax')
return softmax
softmax = inception(21841, 1.0)
batch_size = 64
num_gpu = 4
gpus = [mx.gpu(i) for i in range(num_gpu)]
input_shape = (3, 224, 224)
train = ilsvrc12_iterator(batch_size=batch_size, input_shape=(3,224,224))
model_prefix = "model/Inception-Full"
num_round = 10
logging.info("This script is used to train ImageNet fullset over 21841 classes.")
logging.info("For noraml 1000 classes problem, please use inception.py")
model = mx.model.FeedForward(ctx=gpus, symbol=softmax, num_round=num_round,
learning_rate=0.05, momentum=0.9, wd=0.00001)
model.fit(X=train,
eval_metric="acc",
batch_end_callback=[mx.callback.Speedometer(batch_size), mx.callback.log_train_metric(100)],
epoch_end_callback=mx.callback.do_checkpoint(model_prefix))
# stage 4
flatten = mx.symbol.Flatten(data=pool3)
fc1 = mx.symbol.FullyConnected(data=flatten, num_hidden=4096)
relu6 = mx.symbol.Activation(data=fc1, act_type="relu")
dropout1 = mx.symbol.Dropout(data=relu6, p=0.5)
# stage 5
fc2 = mx.symbol.FullyConnected(data=dropout1, num_hidden=4096)
relu7 = mx.symbol.Activation(data=fc2, act_type="relu")
dropout2 = mx.symbol.Dropout(data=relu7, p=0.5)
# stage 6
fc3 = mx.symbol.FullyConnected(data=dropout2, num_hidden=1000)
softmax = mx.symbol.Softmax(data=fc3)
## data
train, val = ilsvrc12_iterator(batch_size=256, input_shape=(3,224,224))
## train
num_gpus = 2
gpus = [mx.gpu(i) for i in range(num_gpus)]
model = mx.model.FeedForward(
ctx = gpus,
symbol = softmax,
num_round = 20,
learning_rate = 0.01,
momentum = 0.9,
wd = 0.00001)
logging.basicConfig(level = logging.DEBUG)
model.fit(X = train, eval_data = val,
epoch_end_callback = mx.callback.Speedometer(100))
