def __call__(self, iteration, wake_theta_loss, wake_phi_loss, elbo,
generative_model, inference_network, optimizer_theta,
optimizer_phi):
if iteration % self.logging_interval == 0:
util.print_with_time(
'Iteration {} losses: theta = {:.3f}, phi = {:.3f}, elbo = '
'{:.3f}'.format(iteration, wake_theta_loss, wake_phi_loss,
elbo))
self.wake_theta_loss_history.append(wake_theta_loss)
self.wake_phi_loss_history.append(wake_phi_loss)
self.elbo_history.append(elbo)
if iteration % self.checkpoint_interval == 0:
stats_path = util.get_stats_path(self.save_dir)
util.save_object(self, stats_path)
util.save_checkpoint(self.save_dir,
iteration,
generative_model=generative_model,
inference_network=inference_network)
if iteration % self.eval_interval == 0:
log_p, kl = eval_gen_inf(generative_model, inference_network,
self.test_data_loader,
self.eval_num_particles)
self.log_p_history.append(log_p)
self.kl_history.append(kl)
stats = util.OnlineMeanStd()
for _ in range(10):
generative_model.zero_grad()
wake_theta_loss, elbo = losses.get_wake_theta_loss(
generative_model, inference_network, self.test_obs,
self.num_particles)
wake_theta_loss.backward()
theta_grads = [
p.grad.clone() for p in generative_model.parameters()
]
inference_network.zero_grad()
wake_phi_loss = losses.get_wake_phi_loss(
generative_model, inference_network, self.test_obs,
self.num_particles)
wake_phi_loss.backward()
phi_grads = [p.grad for p in inference_network.parameters()]
stats.update(theta_grads + phi_grads)
self.grad_std_history.append(stats.avg_of_means_stds()[1].item())
util.print_with_time(
'Iteration {} log_p = {:.3f}, kl = {:.3f}'.format(
iteration, self.log_p_history[-1], self.kl_history[-1]))
def __call__(self, iteration, loss, elbo, generative_model,
inference_network, optimizer):
if iteration % self.logging_interval == 0:
util.print_with_time(
'Iteration {} loss = {:.3f}, elbo = {:.3f}'.format(
iteration, loss, elbo))
self.loss_history.append(loss)
self.elbo_history.append(elbo)
if iteration % self.checkpoint_interval == 0:
stats_path = util.get_stats_path(self.save_dir)
util.save_object(self, stats_path)
util.save_checkpoint(self.save_dir,
iteration,
generative_model=generative_model,
inference_network=inference_network)
if iteration % self.eval_interval == 0:
log_p, kl = eval_gen_inf(generative_model, inference_network,
self.test_data_loader,
self.eval_num_particles)
_, renyi = eval_gen_inf_alpha(generative_model, inference_network,
self.test_data_loader,
self.eval_num_particles, self.alpha)
self.log_p_history.append(log_p)
self.kl_history.append(kl)
self.renyi_history.append(renyi)
stats = util.OnlineMeanStd()
for _ in range(10):
generative_model.zero_grad()
inference_network.zero_grad()
loss, elbo = losses.get_thermo_alpha_loss(
generative_model, inference_network, self.test_obs,
self.partition, self.num_particles, self.alpha,
self.integration)
loss.backward()
stats.update([p.grad for p in generative_model.parameters()] +
[p.grad for p in inference_network.parameters()])
self.grad_std_history.append(stats.avg_of_means_stds()[1].item())
util.print_with_time(
'Iteration {} log_p = {:.3f}, kl = {:.3f}, renyi = {:.3f}'.
format(iteration, self.log_p_history[-1], self.kl_history[-1],
self.renyi_history[-1]))
def __call__(self, iteration, loss, elbo, generative_model,
inference_network, optimizer):
if iteration % self.logging_interval == 0:
util.print_with_time(
'Iteration {} loss = {:.3f}, elbo = {:.3f}'.format(
iteration, loss, elbo))
self.loss_history.append(loss)
self.elbo_history.append(elbo)
if iteration % self.checkpoint_interval == 0:
stats_filename = util.get_stats_path(self.model_folder)
util.save_object(self, stats_filename)
util.save_models(generative_model, inference_network,
self.model_folder, iteration)
if iteration % self.eval_interval == 0:
self.p_error_history.append(
util.get_p_error(self.true_generative_model, generative_model))
self.q_error_history.append(
util.get_q_error(self.true_generative_model, inference_network,
self.test_obss))
stats = util.OnlineMeanStd()
for _ in range(10):
inference_network.zero_grad()
if self.train_mode == 'vimco':
loss, elbo = losses.get_vimco_loss(generative_model,
inference_network,
self.test_obss,
self.num_particles)
elif self.train_mode == 'reinforce':
loss, elbo = losses.get_reinforce_loss(
generative_model, inference_network, self.test_obss,
self.num_particles)
loss.backward()
stats.update([p.grad for p in inference_network.parameters()])
self.grad_std_history.append(stats.avg_of_means_stds()[1])
util.print_with_time(
'Iteration {} p_error = {:.3f}, q_error_to_true = '
'{:.3f}'.format(iteration, self.p_error_history[-1],
self.q_error_history[-1]))
def get_mean_stds(generative_model, inference_network, num_mc_samples, obss,
num_particles):
vimco_grad = util.OnlineMeanStd()
vimco_one_grad = util.OnlineMeanStd()
reinforce_grad = util.OnlineMeanStd()
reinforce_one_grad = util.OnlineMeanStd()
two_grad = util.OnlineMeanStd()
log_evidence_stats = util.OnlineMeanStd()
log_evidence_grad = util.OnlineMeanStd()
wake_phi_loss_grad = util.OnlineMeanStd()
log_Q_grad = util.OnlineMeanStd()
sleep_loss_grad = util.OnlineMeanStd()
for mc_sample_idx in range(num_mc_samples):
util.print_with_time('MC sample {}'.format(mc_sample_idx))
log_weight, log_q = losses.get_log_weight_and_log_q(
generative_model, inference_network, obss, num_particles)
log_evidence = torch.logsumexp(log_weight, dim=1) - \
np.log(num_particles)
avg_log_evidence = torch.mean(log_evidence)
log_Q = torch.sum(log_q, dim=1)
avg_log_Q = torch.mean(log_Q)
reinforce_one = torch.mean(log_evidence.detach() * log_Q)
reinforce = reinforce_one + avg_log_evidence
vimco_one = 0
for i in range(num_particles):
log_weight_ = log_weight[:, util.range_except(num_particles, i)]
control_variate = torch.logsumexp(
torch.cat([log_weight_, torch.mean(log_weight_, dim=1,
keepdim=True)], dim=1),
dim=1)
vimco_one = vimco_one + (log_evidence.detach() -
control_variate.detach()) * log_q[:, i]
vimco_one = torch.mean(vimco_one)
vimco = vimco_one + avg_log_evidence
normalized_weight = util.exponentiate_and_normalize(log_weight, dim=1)
wake_phi_loss = torch.mean(
-torch.sum(normalized_weight.detach() * log_q, dim=1))
inference_network.zero_grad()
generative_model.zero_grad()
vimco.backward(retain_graph=True)
vimco_grad.update([param.grad for param in
inference_network.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
vimco_one.backward(retain_graph=True)
vimco_one_grad.update([param.grad for param in
inference_network.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
reinforce.backward(retain_graph=True)
reinforce_grad.update([param.grad for param in
inference_network.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
reinforce_one.backward(retain_graph=True)
reinforce_one_grad.update([param.grad for param in
inference_network.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
avg_log_evidence.backward(retain_graph=True)
two_grad.update([param.grad for param in
inference_network.parameters()])
log_evidence_grad.update([param.grad for param in
generative_model.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
wake_phi_loss.backward(retain_graph=True)
wake_phi_loss_grad.update([param.grad for param in
inference_network.parameters()])
inference_network.zero_grad()
generative_model.zero_grad()
avg_log_Q.backward(retain_graph=True)
log_Q_grad.update([param.grad for param in
inference_network.parameters()])
log_evidence_stats.update([avg_log_evidence.unsqueeze(0)])
sleep_loss = losses.get_sleep_loss(
generative_model, inference_network, num_particles * len(obss))
inference_network.zero_grad()
generative_model.zero_grad()
sleep_loss.backward()
sleep_loss_grad.update([param.grad for param in
inference_network.parameters()])
return list(map(
lambda x: x.avg_of_means_stds(),
[vimco_grad, vimco_one_grad, reinforce_grad, reinforce_one_grad,
two_grad, log_evidence_stats, log_evidence_grad, wake_phi_loss_grad,
log_Q_grad, sleep_loss_grad]))
