def main():
enter_debug()
is_net3 = False
with open('task.train.json','r') as fin:
cfgs = json.load(fin)
load_step = cfgs['load_step']
net_name = cfgs['net_name']
if net_name == 'SRNet3':
is_net3 = True
else:
is_net3 = False
if is_net3:
net = nets.SRNet3(filenames=['data.sino8x.json', 'net.srnet1.json'], batch_size=2, low_shape=[16, 80], high_shape=[128, 640], nb_down_sample=2, load_step=load_step)
else:
net = nets.SRNet4(filenames=['data.sino8x.json', 'net.srnet1.json'], batch_size=2, low_shape=[16, 80], high_shape=[128, 640], nb_down_sample=2, load_step=load_step)
# net = nets.SRNet3(filenames=['data.sino8x.json', 'net.srnet1.json'], load_step=-1)
net.init()
# with datasets.SinoShep(filenames='data.sino8x.json') as dataset:
# ss = dataset.sample()
ss = np.load('to_sr.npz')
nb_images = ss['data0'].shape[0]
srs = []
its = []
for i in tqdm(range(nb_images//2)):
# if is_net3:
# feed = ss
# else:
feed = {
'data3': ss['data3'][2*i:2*i+2, :, :, :],
'data': ss['data2'][2*i:2*i+2, :, :, :],
'data2': ss['data2'][2*i:2*i+2, :, :, :],
'data1': ss['data1'][2*i:2*i+2, :, :, :],
'data0': ss['data0'][2*i:2*i+2, :, :, :],
'label': ss['data0'][2*i:2*i+2, :, :, :]}
pred = net.predict(feed)
srs.append(pred['inference'])
its.append(pred['interp'])
print(pred['inference'].shape)
pred_sr = np.concatenate(srs, axis=0)
pred_it = np.concatenate(its, axis=0)
# pred = net.predict(ss)
np.save('sr.npy', pred_sr)
np.save('it.npy', pred_it)
matplotlib.use('agg')
import matplotlib.pyplot as plt
from xlearn.dataset.mnist import MNISTImage, MNIST2
from xlearn.dataset.sinogram import Sinograms
from xlearn.dataset.flickr import Flickr25k
from xlearn.nets.super_resolution import SRNetInterp, SRSimple, SRF3D, SRClassic
from xlearn.utils.general import with_config, empty_list, enter_debug
from xlearn.utils.image import subplot_images
c = dict()
c['is_cl'] = False
enter_debug()
def init():
if c['net'] is not None:
net = define_net()
else:
net = None
if c['dataset'] is not None:
dataset = define_dataset()
else:
dataset = None
return net, dataset
def define_dataset():
def xln(cfg, debug):
if debug:
print("ENTER DEBUG MODE")
enter_debug()
def xln(config, cfg, debug):
config.config = cfg
config.load()
if debug:
print("ENTER DEBUG MODE")
enter_debug()