from sklearn.linear_model import LassoLarsIC
if torch.cuda.is_available():
torch.set_default_tensor_type('torch.cuda.FloatTensor')
np.random.seed(42)
torch.manual_seed(42)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
v = 0.1
A = 1.0
# Making grid
x = np.linspace(-3, 4, 100)
t = np.linspace(0.5, 5.0, 50)
x_grid, t_grid = np.meshgrid(x, t, indexing='ij')
dataset = Dataset(BurgersDelta, v=v, A=A)
noise_range = np.arange(0.0, 1.01, 0.10)
n_runs = 5
for noise_level in noise_range:
for run in np.arange(n_runs):
X_train, y_train = dataset.create_dataset(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1), n_samples=1000, noise=noise_level, random=True, return_idx=False, random_state=run)
estimator = Clustering(estimator=LassoLarsIC(fit_intercept=False))
config = {'n_in': 2, 'hidden_dims': [30, 30, 30, 30, 30], 'n_out': 1, 'library_function':library_1D_in, 'library_args':{'poly_order':2, 'diff_order': 3}, 'sparsity_estimator': estimator}
model = DeepModDynamic(**config)
optimizer = torch.optim.Adam(model.parameters(), betas=(0.99, 0.999), amsgrad=True)
train_dynamic(model, X_train, y_train, optimizer, 10000, log_dir=f'runs/cluster_{noise_level:.2f}_run_{run}/')
requires_grad=True)
y_train = torch.tensor(y_noisy[idx, :][:number_of_samples],
dtype=torch.float32)
estimator = LarsCV(fit_intercept=False)
config = {
'n_in': 2,
'hidden_dims': [20, 20, 20, 20, 20, 20, 20],
'n_out': 1,
'library_function': library_1D_in,
'library_args': {
'poly_order': 1,
'diff_order': 4
},
'sparsity_estimator': estimator
}
model = DeepModDynamic(**config)
optimizer = torch.optim.Adam(model.network_parameters(),
betas=(0.99, 0.99),
amsgrad=True)
train(model,
X_train,
y_train,
optimizer,
100000,
loss_func_args={
'start_sparsity_update': 3000,
'sparsity_update_period': 250
})
torch.set_default_tensor_type('torch.cuda.FloatTensor')
np.random.seed(42)
torch.manual_seed(42)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
v = 0.1
A = 1.0
# Making grid
x = np.linspace(-3, 4, 100)
t = np.linspace(0.5, 5.0, 50)
x_grid, t_grid = np.meshgrid(x, t, indexing='ij')
# Making data
estimator = LassoLarsIC(fit_intercept=False)
dataset = Dataset(BurgersDelta, v=v, A=A)
config = {'n_in': 2, 'hidden_dims': [30, 30, 30, 30, 30], 'n_out': 1, 'library_function':library_1D_in, 'library_args':{'poly_order':2, 'diff_order': 3}, 'sparsity_estimator': estimator}
n_runs = 5
for run_idx in np.arange(n_runs):
X_train, y_train, rand_idx = dataset.create_dataset(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1), n_samples=1000, noise=0.1, random=True, return_idx=True)
theta = dataset.library(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1), poly_order=2, deriv_order=3)[rand_idx, :]
dt = dataset.time_deriv(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1))[rand_idx, :]
model = DeepModDynamic(**config)
optimizer = torch.optim.Adam(model.parameters(), betas=(0.99, 0.999), amsgrad=True)
train(model, X_train, y_train, optimizer, 20000, loss_func_args={'library':torch.tensor(theta) ,'time_deriv': torch.tensor(dt)}, log_dir = f'runs/deepmod_logprob_lstsq_run_{run_idx}')
torch.save(model.state_dict(), f'data/deepmod_logprob_lstsq_run_{run_idx}.pt')
n_samples=1000,
noise=noise_level,
random=True,
return_idx=False,
random_state=run)
estimator = Threshold(threshold=0.1,
estimator=LassoCV(cv=5, fit_intercept=False))
config = {
'n_in': 2,
'hidden_dims': [30, 30, 30, 30, 30],
'n_out': 1,
'library_function': library_1D_in,
'library_args': {
'poly_order': 2,
'diff_order': 3
},
'sparsity_estimator': estimator
}
model = DeepModDynamic(**config)
optimizer = torch.optim.Adam(model.parameters(),
betas=(0.99, 0.999),
amsgrad=True)
print(torch.sum(X_train), torch.sum(y_train))
train_dynamic(model,
X_train,
y_train,
optimizer,
10000,
log_dir=f'runs/threshold_{noise_level:.2f}_run_{run}/')
torch.save(model.state_dict(),
f'data/threshold_model_{noise_level:.2f}_{run}')
random=True)
# Running deepmod
estimator = LassoLarsIC(fit_intercept=False)
config = {
'n_in': 2,
'hidden_dims': [30, 30, 30, 30, 30],
'n_out': 1,
'library_function': library_1D_in,
'library_args': {
'poly_order': 2,
'diff_order': 2
},
'sparsity_estimator': estimator
}
model = DeepModDynamic(**config)
sigma = torch.tensor(0.1, dtype=torch.float32, requires_grad=True)
optimizer = torch.optim.Adam([{
'params': model.network_parameters(),
'betas': (0.99, 0.999)
}, {
'params': sigma,
'betas': (0.99, 0.999)
}],
amsgrad=True)
train_dynamic_logprob(model,
X_train,
y_train,
optimizer,
# Settings and parameters
if torch.cuda.is_available():
torch.set_default_tensor_type('torch.cuda.FloatTensor')
np.random.seed(42)
torch.manual_seed(42)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
v = 0.1
A = 1.0
# Making grid
x = np.linspace(-3, 4, 80)
t = np.linspace(0.5, 5.0, 25)
x_grid, t_grid = np.meshgrid(x, t, indexing='ij')
# Making data
dataset = Dataset(BurgersDelta, v=v, A=A)
X_train, y_train = dataset.create_dataset(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1), n_samples=0, noise=0.1, random=False)
theta = dataset.library(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1), poly_order=2, deriv_order=3)
dt = dataset.time_deriv(x_grid.reshape(-1, 1), t_grid.reshape(-1, 1))
# Running deepmod
config = {'n_in': 2, 'hidden_dims': [30, 30, 30, 30, 30], 'n_out': 1, 'library_function':library_1D_in, 'library_args':{'poly_order':2, 'diff_order': 3}, 'sparsity_estimator': LassoLarsIC(fit_intercept=False)}
model = DeepModDynamic(**config)
optimizer = torch.optim.Adam(model.network_parameters(), betas=(0.99, 0.999), amsgrad=True)
train(model, X_train, y_train, optimizer, 100000, loss_func_args={'library':torch.tensor(theta) ,'time_deriv': torch.tensor(dt)})
torch.save(model.state_dict(), 'data/deepmod_lstsq.pt')