test.py

#!/usr/bin/env python
# encoding: utf-8


""" Unit tests

"""

import numpy
import torch

def test_architectures():

  a = numpy.random.rand(1, 3, 128, 128).astype("float32")
  t = torch.from_numpy(a)

  number_of_classes = 20
  output_dimension = number_of_classes

  # CASIANet
  from ..architectures import CASIANet
  net = CASIANet(number_of_classes)
  embedding_dimension = 320
  output, emdedding = net.forward(t)
  assert output.shape == torch.Size([1, 20])
  assert emdedding.shape == torch.Size([1, 320])

  # CNN8
  from ..architectures import CNN8
  net = CNN8(number_of_classes)
  embedding_dimension = 512
  output, emdedding = net.forward(t)
  assert output.shape == torch.Size([1, 20])
  assert emdedding.shape == torch.Size([1, 512])

  # LightCNN9
  a = numpy.random.rand(1, 1, 128, 128).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import LightCNN9
  net = LightCNN9()
  output, emdedding = net.forward(t)
  assert output.shape == torch.Size([1, 79077])
  assert emdedding.shape == torch.Size([1, 256])

  # LightCNN29
  a = numpy.random.rand(1, 1, 128, 128).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import LightCNN29
  net = LightCNN29()
  output, emdedding = net.forward(t)
  assert output.shape == torch.Size([1, 79077])
  assert emdedding.shape == torch.Size([1, 256])

  # LightCNN29v2
  a = numpy.random.rand(1, 1, 128, 128).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import LightCNN29v2
  net = LightCNN29v2()
  output, emdedding = net.forward(t)
  assert output.shape == torch.Size([1, 79077])
  assert emdedding.shape == torch.Size([1, 256])

  # MCCNN
  a = numpy.random.rand(1, 4, 128, 128).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import MCCNN
  net = MCCNN(num_channels=4)
  output = net.forward(t)
  assert output.shape == torch.Size([1, 1])

  # MCCNNv2
  a = numpy.random.rand(1, 4, 128, 128).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import MCCNNv2
  net = MCCNNv2(num_channels=4)
  output = net.forward(t)
  assert output.shape == torch.Size([1, 1])

  #FASNet 
  a = numpy.random.rand(1, 3, 224, 224).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import FASNet
  net = FASNet(pretrained=False)
  output = net.forward(t)
  assert output.shape == torch.Size([1, 1])

  #DeepMSPAD 
  a = numpy.random.rand(1, 8, 224, 224).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import DeepMSPAD
  net = DeepMSPAD(pretrained=False, num_channels=8)
  output = net.forward(t)
  assert output.shape == torch.Size([1, 1])

  #DeepPixBiS 
  a = numpy.random.rand(1, 3, 224, 224).astype("float32")
  t = torch.from_numpy(a)
  from ..architectures import DeepPixBiS
  net = DeepPixBiS(pretrained=False)
  output = net.forward(t)
  assert output[0].shape == torch.Size([1, 1, 14, 14])
  assert output[1].shape == torch.Size([1, 1])

  # DCGAN
  d = numpy.random.rand(1, 3, 64, 64).astype("float32")
  t = torch.from_numpy(d)
  from ..architectures import DCGAN_discriminator
  discriminator = DCGAN_discriminator(1)
  output = discriminator.forward(t)
  assert output.shape == torch.Size([1])

  g = numpy.random.rand(1, 100, 1, 1).astype("float32")
  t = torch.from_numpy(g)
  from ..architectures import DCGAN_generator
  generator = DCGAN_generator(1)
  output = generator.forward(t)
  assert output.shape == torch.Size([1, 3, 64, 64])

  # Conditional GAN
  d = numpy.random.rand(1, 3, 64, 64).astype("float32")
  t = torch.from_numpy(d)
  cfm = numpy.zeros((1, 13, 64, 64), dtype="float32")
  cfm[:, 0, :, :] = 1
  cfmt = torch.from_numpy(cfm)
  from ..architectures import ConditionalGAN_discriminator
  discriminator = ConditionalGAN_discriminator(13)
  output = discriminator.forward(t, cfmt)
  assert output.shape == torch.Size([1])

  g = numpy.random.rand(1, 100, 1, 1).astype("float32")
  t = torch.from_numpy(g)
  oh = numpy.zeros((1, 13, 1, 1), dtype="float32")
  oh[0] = 1
  oht = torch.from_numpy(oh)
  from ..architectures import ConditionalGAN_generator
  discriminator = ConditionalGAN_generator(100, 13)
  output = discriminator.forward(t, oht)
  assert output.shape == torch.Size([1, 3, 64, 64])


def test_transforms():

  image = numpy.random.rand(3, 128, 128).astype("uint8")

  img = numpy.random.rand(128, 128, 4).astype("uint8")

  from ..datasets import ChannelSelect
  cs = ChannelSelect(selected_channels=[0,1,2])
  assert(cs(img).shape==(128,128,3))

  from ..datasets import RandomHorizontalFlipImage
  rh = RandomHorizontalFlipImage(p=0.5)
  assert(numpy.allclose(rh(rh(img)),img))

  from ..datasets import RollChannels
  sample = {'image': image}
  rc = RollChannels()
  rc(sample)
  assert sample['image'].shape == (128, 128, 3)

  from ..datasets import ToTensor
  tt = ToTensor()
  tt(sample)
  assert isinstance(sample['image'], torch.Tensor)
  # grayscale
  image_gray = numpy.random.rand(128, 128).astype("uint8")
  sample_gray = {'image': image_gray}
  tt(sample_gray)
  assert isinstance(sample['image'], torch.Tensor)

  from ..datasets import Normalize
  image_copy = torch.Tensor(sample['image'])
  norm = Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
  norm(sample)
  for c in range(3):
    for h in range(sample['image'].shape[0]):
      for w in range(sample['image'].shape[1]):
        assert (abs(sample['image'][c, h, w]) - abs((image_copy[c, h, w] - 0.5) / 0.5)) < 1e-10


def test_map_labels():

  labels = ['1', '4', '7']
  from ..datasets import map_labels
  new_labels = map_labels(labels)
  assert '0' in new_labels, "new_labels = {}".format(new_labels)
  assert '1' in new_labels, "new_labels = {}".format(new_labels)
  assert '2' in new_labels, "new_labels = {}".format(new_labels)
  #new_labels = sorted(new_labels)
  #assert new_labels == ['0', '1', '2']

  new_labels = map_labels(labels, start_index = 5)
  #new_labels = sorted(new_labels)
  assert '5' in new_labels, "new_labels = {}".format(new_labels)
  assert '6' in new_labels, "new_labels = {}".format(new_labels)
  assert '7' in new_labels, "new_labels = {}".format(new_labels)
  #assert new_labels == ['5', '6', '7']


from torch.utils.data import Dataset
class DummyDataSet(Dataset):
  def __init__(self):
    pass
  def __len__(self):
    return 100
  def __getitem__(self, idx):
    data =  numpy.random.rand(1, 128, 128).astype("float32")
    label = numpy.random.randint(20)
    sample = {'image': torch.from_numpy(data), 'label': label}
    return sample


def test_CNNtrainer():

  from ..architectures import LightCNN9
  net = LightCNN9(20)

  dataloader = torch.utils.data.DataLoader(DummyDataSet(), batch_size=32, shuffle=True)

  from ..trainers import CNNTrainer
  trainer = CNNTrainer(net, verbosity_level=3)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('model_1_0.pth')

  os.remove('model_1_0.pth')

class DummyDataSetMCCNN(Dataset):
  def __init__(self):
    pass
  def __len__(self):
    return 100
  def __getitem__(self, idx):
    data =  numpy.random.rand(4, 128, 128).astype("float32")
    label = numpy.random.randint(2)
    sample = data, label
    return sample 


def test_MCCNNtrainer():

  from ..architectures import MCCNN
  net = MCCNN(num_channels=4)

  dataloader={}
  dataloader['train'] = torch.utils.data.DataLoader(DummyDataSetMCCNN(), batch_size=32, shuffle=True)

  from ..trainers import MCCNNTrainer
  trainer = MCCNNTrainer(net, verbosity_level=3, do_crossvalidation=False)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('model_1_0.pth')

  os.remove('model_1_0.pth')   


def test_MCCNNtrainer_cv():

  from ..architectures import MCCNN
  net = MCCNN(num_channels=4)

  dataloader={}
  dataloader['train'] = torch.utils.data.DataLoader(DummyDataSetMCCNN(), batch_size=32, shuffle=True)
  dataloader['val'] = torch.utils.data.DataLoader(DummyDataSetMCCNN(), batch_size=32, shuffle=True)

  from ..trainers import MCCNNTrainer
  trainer = MCCNNTrainer(net, verbosity_level=3, do_crossvalidation=True)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('model_1_0.pth')
  assert os.path.isfile('model_100_0.pth') # the best model

  os.remove('model_1_0.pth')  
  os.remove('model_100_0.pth')   


class DummyDataSetFASNet(Dataset):
  def __init__(self):
    pass
  def __len__(self):
    return 100
  def __getitem__(self, idx):
    data =  numpy.random.rand(3, 224,224).astype("float32")
    label = numpy.random.randint(2)
    sample = data, label
    return sample 

def test_FASNettrainer():

  from ..architectures import FASNet
  net = FASNet(pretrained=False)

  dataloader={}
  dataloader['train'] = torch.utils.data.DataLoader(DummyDataSetFASNet(), batch_size=32, shuffle=True)

  from ..trainers import FASNetTrainer
  trainer = FASNetTrainer(net, verbosity_level=3,do_crossvalidation=False)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('model_1_0.pth')

  os.remove('model_1_0.pth') 


def test_FASNettrainer_cv():

  from ..architectures import FASNet
  net = FASNet(pretrained=False)

  dataloader={}
  dataloader['train'] = torch.utils.data.DataLoader(DummyDataSetFASNet(), batch_size=32, shuffle=True)
  dataloader['val'] = torch.utils.data.DataLoader(DummyDataSetFASNet(), batch_size=32, shuffle=True)

  from ..trainers import FASNetTrainer
  trainer = FASNetTrainer(net, verbosity_level=3,do_crossvalidation=True)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('model_1_0.pth')
  assert os.path.isfile('model_100_0.pth')

  os.remove('model_1_0.pth')
  os.remove('model_100_0.pth')

class DummyDataSetGAN(Dataset):
  def __init__(self):
    pass
  def __len__(self):
    return 100
  def __getitem__(self, idx):
    data =  numpy.random.rand(3, 64, 64).astype("float32")
    sample = {'image': torch.from_numpy(data)}
    return sample

def test_DCGANtrainer():

  from ..architectures import DCGAN_generator
  from ..architectures import DCGAN_discriminator
  g = DCGAN_generator(1)
  d = DCGAN_discriminator(1)

  dataloader = torch.utils.data.DataLoader(DummyDataSetGAN(), batch_size=32, shuffle=True)

  from ..trainers import DCGANTrainer
  trainer = DCGANTrainer(g, d, batch_size=32, noise_dim=100, use_gpu=False, verbosity_level=2)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('fake_samples_epoch_000.png')
  assert os.path.isfile('netD_epoch_0.pth')
  assert os.path.isfile('netG_epoch_0.pth')

  os.remove('fake_samples_epoch_000.png')
  os.remove('netD_epoch_0.pth')
  os.remove('netG_epoch_0.pth')

class DummyDataSetConditionalGAN(Dataset):
  def __init__(self):
    pass
  def __len__(self):
    return 100
  def __getitem__(self, idx):
    data =  numpy.random.rand(3, 64, 64).astype("float32")
    sample = {'image': torch.from_numpy(data), 'pose': numpy.random.randint(0, 13)}
    return sample

def test_ConditionalGANTrainer():

  from ..architectures import ConditionalGAN_generator
  from ..architectures import ConditionalGAN_discriminator
  g = ConditionalGAN_generator(100, 13)
  d = ConditionalGAN_discriminator(13)

  dataloader = torch.utils.data.DataLoader(DummyDataSetConditionalGAN(), batch_size=32, shuffle=True)

  from ..trainers import ConditionalGANTrainer
  trainer = ConditionalGANTrainer(g, d, [3, 64, 64], batch_size=32, noise_dim=100, conditional_dim=13)
  trainer.train(dataloader, n_epochs=1, output_dir='.')

  import os
  assert os.path.isfile('fake_samples_epoch_000.png')
  assert os.path.isfile('netD_epoch_0.pth')
  assert os.path.isfile('netG_epoch_0.pth')
  os.remove('fake_samples_epoch_000.png')
  os.remove('netD_epoch_0.pth')
  os.remove('netG_epoch_0.pth')


def test_conv_autoencoder():
    """
    Test the ConvAutoencoder class.
    """
    from bob.learn.pytorch.architectures import ConvAutoencoder

    batch = torch.randn(1, 3, 64, 64)
    model = ConvAutoencoder()
    output = model(batch)
    assert batch.shape == output.shape

    model_embeddings = ConvAutoencoder(return_latent_embedding = True)
    embedding = model_embeddings(batch)
    assert list(embedding.shape) == [1, 16, 5, 5]


def test_extractors():

  # lightCNN9
  from bob.learn.pytorch.extractor.image import LightCNN9Extractor
  extractor = LightCNN9Extractor()
  # this architecture expects 128x128 grayscale images
  data = numpy.random.rand(128, 128).astype("float32")
  output = extractor(data)
  assert output.shape[0] == 256

  # lightCNN29
  from bob.learn.pytorch.extractor.image import LightCNN29Extractor
  extractor = LightCNN29Extractor()
  # this architecture expects 128x128 grayscale images
  data = numpy.random.rand(128, 128).astype("float32")
  output = extractor(data)
  assert output.shape[0] == 256

  # lightCNN29v2
  from bob.learn.pytorch.extractor.image import LightCNN29v2Extractor
  extractor = LightCNN29v2Extractor()
  # this architecture expects 128x128 grayscale images
  data = numpy.random.rand(128, 128).astype("float32")
  output = extractor(data)
  assert output.shape[0] == 256

  # MCCNN 
  from ..extractor.image import MCCNNExtractor
  extractor = MCCNNExtractor(num_channels_used=4)
  # this architecture expects num_channelsx128x128 Multi channel images
  data = numpy.random.rand(4, 128, 128).astype("float32")
  output = extractor(data)
  assert output.shape[0] == 1

  # MCCNNv2
  from ..extractor.image import MCCNNv2Extractor
  extractor = MCCNNv2Extractor(num_channels_used=4)
  # this architecture expects num_channelsx128x128 Multi channel images
  data = numpy.random.rand(4, 128, 128).astype("float32")
  output = extractor(data)
  assert output.shape[0] == 1

  # FASNet
  from ..extractor.image import FASNetExtractor
  extractor = FASNetExtractor()
  # this architecture expects RGB images of size 3x224x224 channel images
  data = numpy.random.rand(3, 224, 224).astype("uint8")
  output = extractor(data)
  assert output.shape[0] == 1