'params': model.gp_layer.variational_parameters(),
'lr': lr * 100
},
{
'params': likelihood.parameters(),
'lr': lr * 100
},
],
lr=lr,
momentum=0.9,
nesterov=True,
weight_decay=0)
mll = mlls.VariationalELBO(
likelihood,
model.gp_layer,
num_data=int(
(len(X_train_abnormal) + len(X_train_normal)) * no_segments))
losses = []
kl_z_loss = []
kl_u_loss = []
exp_log_prob_loss = []
best_auc_eval = 0
best_auc_test = 0
aucs_eval = []
aucs_test = []
X_train = np.concatenate([X_train_abnormal, X_train_normal])
videos = np.concatenate(
[video_names_abnormal_train, video_names_normal_train])
idx = list(range(len(X_train)))
# If you run this example without CUDA, I hope you like waiting!
if torch.cuda.is_available():
model = model.cuda()
likelihood = likelihood.cuda()
n_epochs = 1
lr = 0.1
optimizer = SGD([
{'params': model.feature_extractor.parameters(), 'weight_decay': 1e-4},
{'params': model.gp_layer.hyperparameters(), 'lr': lr * 0.01},
{'params': model.gp_layer.variational_parameters()},
{'params': likelihood.parameters()},
], lr=lr, momentum=0.9, nesterov=True, weight_decay=0)
scheduler = MultiStepLR(optimizer, milestones=[0.5 * n_epochs, 0.75 * n_epochs], gamma=0.1)
mll = mlls.VariationalELBO(likelihood, model.gp_layer, num_data=len(train_loader.dataset))
#
#def train(epoch):
# model.train()
# likelihood.train()
#
# minibatch_iter = tqdm.tqdm(train_loader, desc=f"(Epoch {epoch}) Minibatch")
# with gpytorch.settings.num_likelihood_samples(8):
# for data, target in minibatch_iter:
# if torch.cuda.is_available():
# data, target = data.cuda(), target.cuda()
# optimizer.zero_grad()
# output = model(data)
# loss = -mll(output, target)
# loss.backward()