# Perform some random transformations of the graded potential neuron
# data:
# temp = np.random.randint(0, 5, self.N_in_gpot)
# for i in in_gpot_dict.keys():
# temp += np.random.randint(-1, 1, 1)*in_gpot_dict[i][0]
# out_gpot[:] = temp
out_gpot[:] = np.random.rand(self.N_gpot)
# Randomly select neurons to emit spikes:
# out_spike[:] = \
# sorted(set(np.random.randint(0, self.N_in_spike,
# np.random.randint(0, self.N_in_spike))))
out_spike[:] = np.arange(self.N_spike)
logger = base.setup_logger()
man = Manager(get_random_port(), get_random_port())
man.add_brok()
N1_gpot = N1_spike = 1
N2_gpot = N2_spike = 2
m1 = man.add_mod(MyModule(N1_gpot, N1_spike, man.port_data, man.port_ctrl))
m2 = man.add_mod(MyModule(N2_gpot, N2_spike, man.port_data, man.port_ctrl))
# m3 = MyModule(N, 'unconnected', man.port_data, man.port_ctrl)
# man.add_mod(m3)
# m4 = MyModule(N-2, 'unconnected', man.port_data, man.port_ctrl)
# man.add_mod(m4)
conn1 = Connectivity(N1_gpot, N1_spike, N2_gpot, N2_spike, 1, m1.id, m2.id)
# c1to2['all',:,'all',:,0,'+'] = \
pat12['/a/out/spike0', '/b/in/spike0'] = 1
pat12['/a/out/spike1', '/b/in/spike1'] = 1
pat12['/b/out/spike0', '/a/in/spike0'] = 1
pat12['/b/out/spike1', '/a/in/spike1'] = 1
man.connect(m1, m2, pat12, 0, 1)
# To set the emulation to exit after executing a fixed number of steps,
# start it as follows and remove the sleep statement:
man.start(steps=steps)
# man.start()
# time.sleep(2)
man.stop()
return m1
# Set up logging:
logger = setup_logger(screen=False)
steps = 100
# Emulation 1
start_time = time.time()
size = 2
m1 = emulate(size, steps)
print('Simulation of size {} complete: Duration {} seconds'.format(
size,
time.time() - start_time))
# Emulation 2
start_time = time.time()
size = 100
emulate(size, steps)
print('Simulation of size {} complete: Duration {} seconds'.format(
size,
# Perform some random transformations of the graded potential neuron
# data:
# temp = np.random.randint(0, 5, self.N_in_gpot)
# for i in in_gpot_dict.keys():
# temp += np.random.randint(-1, 1, 1)*in_gpot_dict[i][0]
# out_gpot[:] = temp
out_gpot[:] = np.random.rand(self.N_gpot)
# Randomly select neurons to emit spikes:
# out_spike[:] = \
# sorted(set(np.random.randint(0, self.N_in_spike,
# np.random.randint(0, self.N_in_spike))))
out_spike[:] = np.arange(self.N_spike)
logger = base.setup_logger()
man = Manager(get_random_port(), get_random_port())
man.add_brok()
N1_gpot = N1_spike = 1
N2_gpot = N2_spike = 2
m1 = man.add_mod(MyModule(N1_gpot, N1_spike,
man.port_data, man.port_ctrl))
m2 = man.add_mod(MyModule(N2_gpot, N2_spike,
man.port_data, man.port_ctrl))
# m3 = MyModule(N, 'unconnected', man.port_data, man.port_ctrl)
# man.add_mod(m3)
# m4 = MyModule(N-2, 'unconnected', man.port_data, man.port_ctrl)
# man.add_mod(m4)
pat12['/a/out/spike0', '/b/in/spike0'] = 1
pat12['/a/out/spike1', '/b/in/spike1'] = 1
pat12['/b/out/spike0', '/a/in/spike0'] = 1
pat12['/b/out/spike1', '/a/in/spike1'] = 1
man.connect(m1, m2, pat12, 0, 1)
# To set the emulation to exit after executing a fixed number of steps,
# start it as follows and remove the sleep statement:
man.start(steps=steps)
# man.start()
# time.sleep(2)
man.stop()
return m1
# Set up logging:
logger = setup_logger(screen=False)
steps = 100
# Emulation 1
start_time = time.time()
size = 2
m1 = emulate(size, steps)
print('Simulation of size {} complete: Duration {} seconds'.format(
size, time.time() - start_time))
# Emulation 2
start_time = time.time()
size = 100
emulate(size, steps)
print('Simulation of size {} complete: Duration {} seconds'.format(
size, time.time() - start_time))
logger.info('all done')
import os
import pickle
from importlib import import_module
import attrdict
from setting import CUSTOM_SCR
from base import setup_logger
custom = import_module(CUSTOM_SCR)
logger = setup_logger(name=__name__)
class Dataset:
"""
Data Manager
Parameters
----------
load_set : LoadSet or None
log file path and load function
datas : list of data
data list
"""
def __init__(self, load_set=None, datas=None):
assert load_set is None or datas is None
self.load_set = load_set
self.datas = list()
self.globals = attrdict.AttrDict()
self.param = None