def train_model(N, sparsity):
W = np.random.normal(0, 1, [N, N])
W = W * (RHO / np.max(np.abs(np.linalg.eig(W)[0])))
WI = random.uniform(-TAU, TAU, N)
# make sparse
bef = np.count_nonzero(W)
for i in range(N):
for j in range(N):
if i != j and random.random() < sparsity:
W[i, j] = 0.0
W = W * (RHO / np.max(np.abs(np.linalg.eig(W)[0])))
mc_max = -inf
af = np.count_nonzero(W)
print('sparsity', 1.0 - af / bef)
# W = learn_orthonormal(W, eta)
W = learn_orthogonal(W, eta)
af = np.count_nonzero(W)
print('sparsity', 1.0 - af / bef)
while True:
# W = learn_orthonormal(W, eta)
W = learn_orthogonal(W, eta)
af = np.count_nonzero(W)
print('sparsity', 1.0 - af / bef)
last_mc = mc(WI, W, iters_skip=N, iters_train=10 * N, iters_test=1000)
print(last_mc)
if mc_max < last_mc:
mc_max = last_mc
if mc_max > last_mc:
return mc_max
def train_model(N, eta):
W = np.random.normal(0, 1, [N, N])
W = W * (RHO / np.max(np.abs(np.linalg.eig(W)[0])))
WI = random.uniform(-TAU, TAU, N)
mc_max = -inf
opt_iters = 0
W = learn_orthogonal(W, eta)
while True:
W = learn_orthogonal(W, eta)
last_mc = mc(WI, W, iters_skip=N, iters_train=10*N, iters_test=1000)
print('N={}, eta={}, mc={}'.format(N, eta, last_mc))
if mc_max < last_mc:
mc_max = last_mc
if mc_max > last_mc:
print()
return mc_max, opt_iters
# W = learn_orthonormal(W, eta)
opt_iters += 1
def train_model(N, sparsity):
W = np.random.normal(0, 1, [N, N])
W = W * (RHO / np.max(np.abs(np.linalg.eig(W)[0])))
WI = random.uniform(-TAU, TAU, N)
# make sparse
for i in range(N):
for j in range(N):
if i != j and random.random() < sparsity:
W[i, j] = 0.0
W = W * (RHO / np.max(np.abs(np.linalg.eig(W)[0])))
mc_max = -inf
W = learn_orthogonal(W, eta)
while True:
W = learn_orthogonal(W, eta)
last_mc = mc(WI, W, iters_skip=N, iters_train=10*N, iters_test=1000)
if mc_max < last_mc:
mc_max = last_mc
if mc_max > last_mc:
return mc_max
iterations=iterations,
iterations_coef_measure=1000,
use_input=False,
target_later=True,
calc_lyapunov=False)
mc_mean = np.zeros([len(reservoir_sizes), len(taus)])
mc_std = np.zeros([len(reservoir_sizes), len(taus)])
for tau_idx, tau in enumerate(taus):
print(tau_idx)
for rsi, N in enumerate(reservoir_sizes):
mc = np.zeros(INSTANCES)
for inst in range(INSTANCES):
W = random.normal(0, 1, [N, N])
W = W * (rho / np.max(np.abs(np.linalg.eig(W)[0])))
WI = random.uniform(-tau, tau, N)
for _ in range(ORTHOPROCESS_ITERATIONS):
W = learn_orthogonal(W, eta)
mc[inst], _ = measure_mc(W, WI, iterations=10 * N)
print(N, mc[inst])
mc_mean[rsi, tau_idx] = np.average(mc)
mc_std[rsi, tau_idx] = np.std(mc)
np.save('mcm', mc_mean)
np.save('mcs', mc_std)
memory_max=int(1.1*WI.shape[0]),
iterations=1200,
iterations_coef_measure=1000,
use_input=False,
target_later=True,
calc_lyapunov=False)
mc_mean = np.zeros([len(reservoir_sizes), len(taus)])
mc_std = np.zeros([len(reservoir_sizes), len(taus)])
for tau_idx, tau in enumerate(taus):
print(tau_idx)
for rsi, N in enumerate(reservoir_sizes):
mc = np.zeros(INSTANCES)
for inst in range(INSTANCES):
W = random.normal(0, 1, [N, N])
W = W * (rho / np.max(np.abs(np.linalg.eig(W)[0])))
WI = random.uniform(-tau, tau, N)
for _ in range(ORTHOPROCESS_ITERATIONS):
W = learn_orthogonal(W, eta)
mc[inst], _ = measure_mc(W, WI)
mc_mean[rsi, tau_idx] = np.average(mc)
mc_std[rsi, tau_idx] = np.std(mc)
np.save('mcm', mc_mean)
np.save('mcs', mc_std)
WI = np.random.uniform(-tau, tau, N)
WG = W
mc_gon = np.zeros(ORTHO_ITERATIONS + 1)
les = []
mc_gon[0], le = measure_mc(WG, WI, True)
les.append(le)
eigen_values = np.zeros([ORTHO_ITERATIONS + 1, N])
singular_values = np.zeros([ORTHO_ITERATIONS + 1, N])
eigen_values[0, :] = np.sort(np.abs(np.linalg.eig(W)[0]))
singular_values[0, :] = np.linalg.svd(W, compute_uv=False)
for it in range(ORTHO_ITERATIONS):
WG = learn_orthogonal(WG, eta)
mc_gon[it + 1], le = measure_mc(WG, WI, True)
les.append(le)
eigen_values[it + 1, :] = np.sort(np.abs(np.linalg.eig(WG)[0]))
singular_values[it + 1, :] = np.linalg.svd(WG, compute_uv=False)
MCs.append(mc_gon)
EVs.append(eigen_values)
SVs.append(singular_values)
LEs.append(les)
print()
np.save('mc', MCs)
np.save('ev', EVs)
eigenvaluesG = np.zeros([ORTHO_ITERATIONS + 1, N])
singular_valuesG = np.zeros([ORTHO_ITERATIONS + 1, N])
eigenvaluesN = np.zeros([ORTHO_ITERATIONS + 1, N])
singular_valuesN = np.zeros([ORTHO_ITERATIONS + 1, N])
mc_gon[0] = measure_mc(WG, WI)
mc_nor[0] = measure_mc(WN, WI)
eigenvaluesG[0, :] = np.sort(np.abs(np.linalg.eig(W)[0]))
singular_valuesG[0, :] = np.linalg.svd(W, compute_uv=False)
eigenvaluesN[0, :] = np.sort(np.abs(np.linalg.eig(W)[0]))
singular_valuesN[0, :] = np.linalg.svd(W, compute_uv=False)
for it in range(ORTHO_ITERATIONS):
print('\riteration', it, 'of', ORTHO_ITERATIONS, end='')
WG = learn_orthogonal(WG, eta)
WN = learn_orthonormal(WN, eta)
mc_gon[it + 1] = measure_mc(WG, WI)
mc_nor[it + 1] = measure_mc(WN, WI)
eigenvaluesG[it + 1, :] = np.sort(np.abs(np.linalg.eig(WG)[0]))
singular_valuesG[it + 1, :] = np.linalg.svd(WG, compute_uv=False)
eigenvaluesN[it + 1, :] = np.sort(np.abs(np.linalg.eig(WN)[0]))
singular_valuesN[it + 1, :] = np.linalg.svd(WN, compute_uv=False)
print()
def replot():