def train(param=PARAMS, sv=SOLVE, small=False):
sv['name'] = 'TEST'
input_var = raw_input('Are you testing now? ')
if 'no' in input_var:
sv.pop('name')
else:
sv['name'] += input_var
#out = u.get(6,small=True, aug=True)
imgs, ll = load_rnn_pk(files)
imgs = imgs.reshape((-1,1,256,256))
ll = ll.reshape((-1,1,256,256))
datas = u.prepare_set(imgs, ll)
out = u.create_iter(*datas, batch_size=5)
net = cnn_net(
use_logis=True
)
param['eval_data'] = out[1]
s = Solver(net, out[0], sv, **param)
s.train()
s.predict()
s.all_to_png()
s.save_best_model()
s.plot_process()
def main(argv):
parser = argparse.ArgumentParser()
# Optional arguments.
parser.add_argument(
"--epochs",
default=200,
help="The number of train iterations",
)
parser.add_argument(
"--batch_size_train",
default=100,
help="number of train samples per batch",
)
parser.add_argument(
"--batch_size_val",
default=40,
help="number of validation samples per batch",
)
parser.add_argument(
"--images_dim",
default=64,
help="'height, width' dimensions of input images.",
)
args = parser.parse_args()
model = cnn_net(args.images_dim)
print("Plotting the model")
plot(model, to_file='model.png')
train(model, args.epochs, args.batch_size_train, args.batch_size_val, args.images_dim)
def main():
net = cnn_net()
img, ll = u.load_pk('../DATA/PK/o1.pk')
ival, lval = u.augment_sunny(img[:5], ll[:5])
val = mx.io.NDArrayIter(ival, label=lval)
model = mx.model.FeedForward.load(
*Aug40,
ctx=u.gpu(1),
learning_rate=6,
num_epoch=10,
optimizer='sgd',
initializer=mx.initializer.Xavier(rnd_type='gaussian'))
u.predict_draw(model, val, folder='MoveCheck')
def train(param = PARAMS, sv=SOLVE, small=False):
sv['name'] = 'TEST'
input_var = raw_input('Are you testing now? ')
if 'no' in input_var:
sv.pop('name')
else:
sv['name'] += input_var
out = u.get(6,small=True, aug=True)
net = cnn_net(
use_logis=False
)
param['eval_data'] = out['val']
s = Solver(net, out['train'], sv, **param)
s.train()
s.predict()
def test():
net = cnn_net()
data = get(5, small = True)
train = data['train']
val = data['val']
for l in [ 1e-3, 3e-3, 6e-3, 1e-2, 3e-2, 6e-2, 1e-1, 3e-1, 6e-1, 1, 3, 6]:
logging.info('------------------------------------%f-------------------------------', l)
c = Callback(name=str(l))
model = mx.model.FeedForward.create(
net,
train,
learning_rate = l,
ctx = [mx.context.gpu(i) for i in [0,1,2]],
eval_data = val,
eval_metric = mx.metric.create(c.eval),
num_epoch = 40,
)
c.all_to_png()
predict_test(model, val, c.path)
import ipt import mxnet as mx import matplotlib.pyplot as plt from cnn_internal import fetch_internal from cnn import cnn_net import my_utils as mu import os p = '/home/zijia/HeartDeepLearning/CNN/Result/<0Save>/<1-17:12:45>[E40]/[ACC-0.92900 E39]' e = 39 net = cnn_net() val = mu.get(2, small=True)['val'] outputs, imgs, lls = fetch_internal(net, val, p, e) stamp = 'Inspect/' + mu.parse_time() + '/' os.makedirs(stamp) mu.save_img(imgs[0, 0], stamp + 'Input') #for k,v in outputs.items(): # if len(v.shape)>2: # mu.save_img(v[0,0], stamp+k)
import ipt import mxnet as mx import matplotlib.pyplot as plt from cnn_internal import fetch_internal from cnn import cnn_net import my_utils as mu import os p = '/home/zijia/HeartDeepLearning/CNN/Result/<0Save>/<1-17:12:45>[E40]/[ACC-0.92900 E39]' e = 39 net = cnn_net() val = mu.get(2,small=True)['val'] outputs, imgs, lls = fetch_internal(net,val,p,e) stamp = 'Inspect/'+mu.parse_time()+'/' os.makedirs(stamp) mu.save_img(imgs[0,0],stamp+'Input') #for k,v in outputs.items(): # if len(v.shape)>2: # mu.save_img(v[0,0], stamp+k)