def generate_predicted_signal(self,
signal2model,
N=2000,
starting_signal=[0],
window_seen_by_GRU_size=256):
signal2model.model_type = ModelType.SGD
model = LibphysSGDGRU.LibphysSGDGRU(signal2model)
model.load(dir_name=signal2model.signal_directory,
file_tag=self.get_file_tag(-5, -5))
return model.generate_predicted_signal(N, starting_signal,
window_seen_by_GRU_size)
def generate_predicted_signal(self,
N=2000,
starting_signal=[0],
window_seen_by_GRU_size=256):
signal2model = Signal2Model(self.model_name,
self.get_directory_tag(),
signal_dim=self.signal_dim,
hidden_dim=self.hidden_dim,
signal_type=self.signal_type,
model_type=ModelType.SGD)
model = LibphysSGDGRU.LibphysSGDGRU(signal2model)
model.load(self.model_name, filetag=self.get_file_tag(-5, -5))
return model.generate_predicted_signal(N, starting_signal,
window_seen_by_GRU_size)
def create_history(subject):
model_info = db.ecg_64_models[subject]
signal2Model = Signal2Model(model_info.dataset_name,
model_info.directory,
signal_dim=model_info.Sd,
hidden_dim=model_info.Hd)
model = GRU.LibphysSGDGRU(signal2Model)
# plt.ion()
# font = {'family': 'lato',
# 'weight': 'bold',
# 'size': 40}
#
# matplotlib.rc('font', **font)
x = [0]
i = 0
window_size = 512
fz = 250
N = 3000
titles = []
signals = []
for epoch in [-1, 20, 30, 50, 80, 410]:
signal = [np.random.randint(0, 63, size=1)[0]]
if epoch < 0:
file_tag = model.get_file_tag(epoch, epoch)
else:
file_tag = model.get_file_tag(0, epoch)
model.load(file_tag=file_tag, dir_name=model_info.directory)
print("Processing epoch " + str(epoch))
for i in range(N):
# print(i)
sample, _ = model.generate_online_predicted_signal(
signal, window_size, uncertaintly=0.01)
signal.append(sample)
x.append((i + 1) / fz)
titles.append("Epoch {0}".format(epoch))
signals.append(np.array(signal))
np.savez("../data/history_of_{0}.npz".format(subject),
titles=titles,
signals=signals,
t=x)
return [titles, signals, x]
import DeepLibphys.utils.functions.database as db
from DeepLibphys.utils.functions.common import *
import DeepLibphys.models.LibphysSGDGRU as GRU
model_info = db.ecg_1024_256_RAW[6]
model = GRU.LibphysSGDGRU(model_info.to_signal2model())
model.load(dir_name=model_info.directory)
model.online_sinthesizer(20000, [60],
window_seen_by_GRU_size=512,
uncertaintly=0.1)
filtered_signals = quantize_signal(filtered_signals, 256)
save_directory = "bento"
sig_model = Signal2Model("GRU_Filter_128",
save_directory,
signal_dim=256,
mini_batch_size=32,
window_size=512)
# maybe change mbs to 16, tune overlap
#print("Training...")
#train_block([original_signals], [filtered_signals], overlap=0.2, n_for_each=128, signal2model=sig_model)
test_signals = loadmat(get_fantasia_full_paths()[0])['val'][0][30000:60000]
test_signals = quantize_signal(test_signals, 256)
model = LibphysSGDGRU.LibphysSGDGRU(sig_model)
model.load(model.get_file_tag(-5, -5), sig_model.signal_directory)
predicted = generate_predicted_signal(model,
3000,
starting_signal=test_signals,
window_seen_by_GRU_size=512)
plt.subplot(311)
plt.plot(original_signals)
plt.subplot(312)
plt.plot(filtered_signals)
plt.subplot(313)
plt.plot(predicted)
plt.show()
signal_directory = 'DAY_HRV_LOO_HF_['+str(hidden_dim)+'.'+str(window_size)+']'
z = 0
for i, group_signals in zip(range(z, len(all_signals)), all_signals[z:]):
model_name = 'day_hrv_rr_{0}_{1}'.format(i, "-1")
signal2model = Signal2Model(model_name, signal_directory,
signal_dim=signal_dim,
window_size=window_size,
hidden_dim=hidden_dim,
mini_batch_size=mini_batch_size,
learning_rate_val=0.1,
save_interval=100000)
if i == 0:
pass
else:
model = GRU.LibphysSGDGRU(signal2model)
model.train_block(group_signals, signal2model, n_for_each=1)
for j in range(len(group_signals)):
if i == 0 and j < 29:
pass
else:
train_group = group_signals
train_group.pop(j)
model_name = 'day_hrv_rr_{0}_{1}'.format(i, j)
signal2model.model_name = model_name
running_ok = False
while not running_ok:
model = GRU.LibphysSGDGRU(signal2model)
running_ok = model.train_block(train_group, signal2model, n_for_each=1)