input_channel = 1
conv_architecture=[input_channel,12,24,48]
#specify encoder
#with convolutional layers and hidden dense layer
encoder = Encoder([ConvLayers(input_dim=input_dim,conv_architecture=conv_architecture, conv_kernel=[4,4,4], conv_stride=[2,2,2], last_activation="sigmoid"),
DenseLayers(architecture=[],output_size=latent_dim)])
#specify decoder-mu
decoder_mu = infer_decoder(encoder,last_activation="sigmoid") #symmetrical to encoder
#combine them into autoencoder
autoencoder = Autoencoder(encoder, decoder_mu)
#convert into BAE-Ensemble
bae_ensemble = BAE_Ensemble(autoencoder=autoencoder, use_cuda=use_cuda, anchored=True, weight_decay=1, num_samples=5)
#train mu network
run_auto_lr_range(train_loader, bae_ensemble)
bae_ensemble.fit(train_loader,num_epochs=1)
#for each model, evaluate and plot:
bae_models = [bae_ensemble]
id_data_test = test_loader
ood_data_list = [ood_loader]
train_set_name = "FashionMNIST"
#run evaluation test of model on ood data set
run_test_model(bae_models=bae_models, id_data_test=test_loader, ood_data_list=ood_data_list, id_data_name=train_set_name, output_reshape_size=(28, 28))
last_activation="sigmoid") #symmetrical to encoder
#combine them into autoencoder
autoencoder = Autoencoder(encoder, decoder_mu)
#convert into VAE
bae_model = VAE(
autoencoder=autoencoder,
num_train_samples=5,
num_samples=25, #during prediction only
use_cuda=use_cuda,
weight_decay=0.001,
likelihood="bernoulli")
#train mu network
run_auto_lr_range(train_loader, bae_model)
bae_model.fit(train_loader, num_epochs=10)
#for each model, evaluate and plot:
bae_models = [bae_model]
id_data_test = test_loader
ood_data_list = [ood_loader]
train_set_name = "CIFAR"
#run evaluation test of model on ood data set
run_test_model(bae_models=bae_models,
id_data_test=test_loader,
ood_data_list=ood_data_list,
id_data_name=train_set_name,
output_reshape_size=(32, 32, 3))
#specify decoder-mu
decoder_mu = infer_decoder(encoder,
last_activation="sigmoid") #symmetrical to encoder
#combine them into autoencoder
autoencoder = Autoencoder(encoder, decoder_mu)
#convert into BAE-MCDropout
bae_mcdropout = BAE_MCDropout(autoencoder=autoencoder,
dropout_p=0.2,
num_train_samples=5,
num_samples=50,
use_cuda=use_cuda)
#train mu network
run_auto_lr_range(train_loader, bae_mcdropout)
bae_mcdropout.fit(train_loader, num_epochs=1)
#for each model, evaluate and plot:
bae_models = [bae_mcdropout]
id_data_test = test_loader
ood_data_list = [ood_loader]
train_set_name = "FashionMNIST"
#run evaluation test of model on ood data set
run_test_model(bae_models=bae_models,
id_data_test=test_loader,
ood_data_list=ood_data_list,
id_data_name=train_set_name,
output_reshape_size=(28, 28))