def plot_compare_spec_results2(model, proc):
import cPickle
from smartNN.utils.utils import tile_raster_graphs
from PIL.Image import fromarray
import smartNN.datasets.preprocessor as processor
with open(os.environ['smartNN_SAVE_PATH'] + '/log/' + model + '/model.pkl', 'rb') as f:
mlp = cPickle.load(f)
start = -12
end = -10
tile_shape=(2,1)
print('..loading data from spec.double')
folder_path = os.environ['smartNN_DATA_PATH'] + '/p276_double'
dct_data = extract_examples(folder_path, -10, -1)
dct_data = dct_data[start:end]
print('..loading data from ' + model)
data = P276(train_valid_test_ratio=[8,1,1])
test = data.get_test()
# prep = Standardize()
# prep = Scale()
prep = getattr(processor, proc)()
if prep.__class__.__name__ == 'Scale':
prep.max = 98803.031
prep.min = 0.1
print('..preprocessing data ' + prep.__class__.__name__ )
proc_test_X = prep.apply(test.X)
dct_data = prep.apply(dct_data)
print('..fprop X')
output = mlp.fprop(proc_test_X)
print('..saving data')
plt = tile_raster_graphs(dct_data, proc_test_X[start:end], output[start:end], slice=(0,-1),
tile_shape=tile_shape, tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_0_2049_all2.png')
plt.close()
plt = tile_raster_graphs(dct_data, proc_test_X[-12:-10], output[-12:-10], slice=(0,200),
tile_shape=(2,1), tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_0_200_all2.png')
plt.close()
plt = tile_raster_graphs(dct_data, proc_test_X[-12:-10], output[-12:-10], slice=(1500,-1),
tile_shape=(2,1), tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_1500_2400_all2.png')
plt.close()
print('Saved Successfully')
def plot_compare_spec_results(model, proc):
import cPickle
from smartNN.utils.utils import tile_raster_graphs
from PIL.Image import fromarray
import smartNN.datasets.preprocessor as processor
with open(os.environ['smartNN_SAVE_PATH'] + '/log/' + model + '/model.pkl', 'rb') as f:
mlp = cPickle.load(f)
# data = Mnist(preprocessor = None,
# binarize = False,
# batch_size = 100,
# num_batches = None,
# train_ratio = 5,
# valid_ratio = 1,
# iter_class = 'SequentialSubsetIterator',
# rng = None)
start = -12
end = -10
tile_shape=(2,1)
print('..loading data from ' + model)
data = P276(train_valid_test_ratio=[8,1,1])
test = data.get_test()
# prep = Standardize()
# prep = Scale()
prep = getattr(processor, proc)()
if prep.__class__.__name__ == 'Scale':
prep.max = 98803.031
prep.min = 0.1
print('..preprocessing data ' + prep.__class__.__name__ )
proc_test_X = prep.apply(test.X)
print('..fprop X')
output = mlp.fprop(proc_test_X)
print('..saving data')
plt = tile_raster_graphs(proc_test_X[start:end], output[start:end], slice=(0,-1),
tile_shape=tile_shape, tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_0_2049.png')
plt.close()
plt = tile_raster_graphs(proc_test_X[-12:-10], output[-12:-10], slice=(0,200),
tile_shape=(2,1), tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_0_200.png')
plt.close()
plt = tile_raster_graphs(proc_test_X[-12:-10], output[-12:-10], slice=(1500,-1),
tile_shape=(2,1), tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model + '_1500_2400.png')
plt.close()
print('Saved Successfully')
def plot_spec(spec):
with open(spec) as f:
spec_data = np.fromfile(f, dtype='<f4', count=-1)
plt = tile_raster_graphs(spec_data, spec_data, slice=(0,-1),
tile_shape=(2,1), tile_spacing=(0.1,0.1), legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/spec.png')
plt.show()
plt.close()
def plot_spec(spec):
with open(spec) as f:
spec_data = np.fromfile(f, dtype='<f4', count=-1)
plt = tile_raster_graphs(spec_data,
spec_data,
slice=(0, -1),
tile_shape=(2, 1),
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/spec.png')
plt.show()
plt.close()
def plot_compare_spec_results(model, proc):
import cPickle
from smartNN.utils.utils import tile_raster_graphs
from PIL.Image import fromarray
import smartNN.datasets.preprocessor as processor
with open(os.environ['smartNN_SAVE_PATH'] + '/log/' + model + '/model.pkl',
'rb') as f:
mlp = cPickle.load(f)
# data = Mnist(preprocessor = None,
# binarize = False,
# batch_size = 100,
# num_batches = None,
# train_ratio = 5,
# valid_ratio = 1,
# iter_class = 'SequentialSubsetIterator',
# rng = None)
start = -12
end = -10
tile_shape = (2, 1)
print('..loading data from ' + model)
data = P276(train_valid_test_ratio=[8, 1, 1])
test = data.get_test()
# prep = Standardize()
# prep = Scale()
prep = getattr(processor, proc)()
if prep.__class__.__name__ == 'Scale':
prep.max = 98803.031
prep.min = 0.1
print('..preprocessing data ' + prep.__class__.__name__)
proc_test_X = prep.apply(test.X)
print('..fprop X')
output = mlp.fprop(proc_test_X)
print('..saving data')
plt = tile_raster_graphs(proc_test_X[start:end],
output[start:end],
slice=(0, -1),
tile_shape=tile_shape,
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_0_2049.png')
plt.close()
plt = tile_raster_graphs(proc_test_X[-12:-10],
output[-12:-10],
slice=(0, 200),
tile_shape=(2, 1),
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_0_200.png')
plt.close()
plt = tile_raster_graphs(proc_test_X[-12:-10],
output[-12:-10],
slice=(1500, -1),
tile_shape=(2, 1),
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_1500_2400.png')
plt.close()
print('Saved Successfully')
def plot_compare_spec_results2(model, proc):
import cPickle
from smartNN.utils.utils import tile_raster_graphs
from PIL.Image import fromarray
import smartNN.datasets.preprocessor as processor
with open(os.environ['smartNN_SAVE_PATH'] + '/log/' + model + '/model.pkl',
'rb') as f:
mlp = cPickle.load(f)
start = -12
end = -10
tile_shape = (2, 1)
print('..loading data from spec.double')
folder_path = os.environ['smartNN_DATA_PATH'] + '/p276_double'
dct_data = extract_examples(folder_path, -10, -1)
dct_data = dct_data[start:end]
print('..loading data from ' + model)
data = P276(train_valid_test_ratio=[8, 1, 1])
test = data.get_test()
# prep = Standardize()
# prep = Scale()
prep = getattr(processor, proc)()
if prep.__class__.__name__ == 'Scale':
prep.max = 98803.031
prep.min = 0.1
print('..preprocessing data ' + prep.__class__.__name__)
proc_test_X = prep.apply(test.X)
dct_data = prep.apply(dct_data)
print('..fprop X')
output = mlp.fprop(proc_test_X)
print('..saving data')
plt = tile_raster_graphs(dct_data,
proc_test_X[start:end],
output[start:end],
slice=(0, -1),
tile_shape=tile_shape,
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_0_2049_all2.png')
plt.close()
plt = tile_raster_graphs(dct_data,
proc_test_X[-12:-10],
output[-12:-10],
slice=(0, 200),
tile_shape=(2, 1),
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_0_200_all2.png')
plt.close()
plt = tile_raster_graphs(dct_data,
proc_test_X[-12:-10],
output[-12:-10],
slice=(1500, -1),
tile_shape=(2, 1),
tile_spacing=(0.1, 0.1),
legend=True)
plt.savefig(os.environ['smartNN_SAVE_PATH'] + '/images/' + model +
'_1500_2400_all2.png')
plt.close()
print('Saved Successfully')
