losses['dis'].append(epochDisLoss / i)
#### Test
dae.eval()
dis.eval()
#get test outuputs and losses
xTest, yTest = prep_data(iter(testLoader).next(), useCUDA=dae.useCUDA)
zTest, recTest = dae.forward(xTest) #N.B. corruption in here
recLossTest = dae.rec_loss(recTest, xTest)
#Plot losses
losses['test rec'].append(recLossTest.data[0])
if e > 0: #only one point for test rec otherwise
plot_losses(losses, exDir, epochs=e + 1)
plot_norm_losses(losses, exDir)
#save parameters
dae.save_params(exDir)
dis.save_params(exDir)
#Save images of original and rec
save_image(xTest.data, join(exDir, 'original.png'))
save_image(recTest.data, join(exDir, 'rec.png'))
#Save samples
sampleDir = join(exDir, 'epoch_' + str(e))
os.mkdir(sampleDir)
print 'sample dir:', sampleDir
dae.sample_x(opts.M, sampleDir)
bceLossTest, klLossTest = vae_loss_fn(rec_x=outputs.data,
x=xTest,
mu=outMu,
logVar=outLogVar)
maxVal, predLabel = torch.max(outY, 1)
classScoreTest = torch.eq(predLabel,
yTest).float().sum() / yTest.size(0)
print 'classification test:', classScoreTest.data[0]
save_image(xTest, join(exDir, 'input.png'))
save_image(outputs.data, join(exDir, 'output_' + str(e) + '.png'))
label_switch(xTest, yTest, cvae, exDir=exDir)
cvae.save_params(exDir=exDir)
losses['total'].append(epochLoss / Ns)
losses['kl'].append(epochLoss_kl / Ns)
losses['bce'].append(epochLoss_bce / Ns)
losses['test_bce'].append(
(bceLossTest).data[0] / xTest.size(0)) #append every epoch
losses['dis'].append(epochLoss_dis / Ns)
losses['gen'].append(epochLoss_gen / Ns)
losses['class'].append(epochLoss_class / Ns)
losses['test_class'].append(classScoreTest.data[0])
if e > 1:
plot_losses(losses, exDir, epochs=e + 1)
plot_norm_losses(losses, exDir, epochs=1 + e)
xTest, yTest = prep_data(testData, classer.useCUDA)
####### eval VAE #######
recTest, outMu, outLogVar = vae(xTest)
bceLossTest, klLossTest = vae.loss(recTest, xTest, outMu, outLogVar)
save_image(xTest.data, join(exDir, 'input.png'))
save_image(recTest.data, join(exDir, 'output_' + str(e) + '.png'))
####### eval CLASSER #######
yPredTrain = classer(xTrain)
yPredTest = classer(xTest)
classTrain = score(yPredTrain, yTrain).data[0]
classTest = score(yPredTest, yTest).data[0]
vaeLosses['test_bce'].append(bceLossTest.mean().data[0])
vaeLosses['test_kl'].append(bceLossTest.mean().data[0])
classerLosses['train_acc'].append(classTrain)
classerLosses['test_acc'].append(classTest)
viz_DZ(deltaZ.data[0].cpu().numpy(), exDir)
if e > 0:
plot_norm_losses(vaeLosses, exDir, e, title='VAE')
plot_losses(classerLosses, exDir, e, title='CLASSER')
plot_losses(deltaZLosses, exDir, e, title='DELTAZ')
#Evaluate the results:
eval_results(vae, classer, deltaZ, testLoader, exDir)
inception_score(vae, classer, deltaZ, testLoader, exDir)
(e, i, epochLoss/i, epochLoss_dis/i, epochLoss_gen/i, epochLoss_bce/i, epochLoss_kl/i, epochLoss_aux/i, time() - TIME)
#generate samples after each 10 epochs
normbceLossTest, classScoreTest = evaluate(cvae,
testLoader,
exDir,
e=e)
cvae.save_params(exDir=exDir)
losses['total'].append(epochLoss / Ns)
losses['kl'].append(epochLoss_kl / Ns)
losses['bce'].append(epochLoss_bce / Ns)
losses['test_bce'].append(normbceLossTest) #append every epoch
losses['dis'].append(epochLoss_dis / Ns)
losses['gen'].append(epochLoss_gen / Ns)
losses['class'].append(epochLoss_class / Ns)
losses['test_class'].append(classScoreTest)
losses['aux'].append(epochLoss_aux / Ns)
losses['auxEnc'].append(epochLoss_auxEnc / Ns)
if e > 1:
plot_losses(losses, exDir, epochs=e + 1)
plot_norm_losses(losses, exDir, epochs=e + 1)
normbceLossTest, classScoreTest = evaluate(cvae,
testLoader,
exDir,
e='evalMode')
