weighted random sampler

######################################################################
# Load Data
# ---------
#
# We will use torchvision and torch.utils.data packages for loading the
# data.
#
# The problem we're going to solve today is to train a model to classify
# **ants** and **bees**. We have about 120 training images each for ants and bees.
# There are 75 validation images for each class. Usually, this is a very
# small dataset to generalize upon, if trained from scratch. Since we
# are using transfer learning, we should be able to generalize reasonably
# well.
#
# This dataset is a very small subset of imagenet.
#
# .. Note ::
#    Download the data from
#    `here <https://download.pytorch.org/tutorial/hymenoptera_data.zip>`_
#    and extract it to the current directory.

# Data augmentation and normalization for training
# Just normalization for validation


#using the weight sampling method from https://github.com/ptrblck/pytorch_misc/blob/master/weighted_sampling.py#L25


data_transforms = {
    'train': transforms.Compose([
        transforms.RandomResizedCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.RandomRotation(20),
        transforms.ColorJitter(0.3, 0.3, 0.3),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
#    'val': transforms.Compose([
#        transforms.Resize(256),
#        transforms.CenterCrop(224),
#        transforms.ToTensor(),
#        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
#    ]),

        'test': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
}


data_dir = "10folds/10fold_9"

class MonaDataset(datasets.folder.ImageFolder):
    def __init__(self, root, transform=None, target_transform=None,
                 loader=datasets.folder.default_loader):
        super(MonaDataset, self).__init__(root, transform, target_transform, loader)

    def __getitem__(self, index):
        path, target = self.samples[index]
        sample = self.loader(path)
        if self.transform is not None:
            sample = self.transform(sample)
        if self.target_transform is not None:
            target = self.target_transform(target)
        return sample, target, path


image_datasets = {x: MonaDataset(os.path.join(data_dir, x),
                                          data_transforms[x])
                  for x in ['train', 'test']}
dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,
                                             shuffle=True, num_workers=4)
              for x in ['train', 'test']}



targets = []
with torch.no_grad():
    for i, (inputs, classes, im_path) in enumerate(dataloaders['train']):
        targets.append(classes.cpu().detach().numpy().tolist())

with torch.no_grad():
    for i, (inputs, classes, im_path) in enumerate(dataloaders['test']):
        targets.append(classes.cpu().detach().numpy().tolist())

flat_targets = [item for sublist in targets for item in sublist]
flat_targets = torch.Tensor(flat_targets)
print(flat_targets)
class_sample_count = torch.tensor([(flat_targets == t).sum() for t in torch.unique(flat_targets, sorted=True)])
print(class_sample_count)
weight = 1. / class_sample_count.float()
print(weight)
samples_weight = torch.tensor([weight[t] for t in flat_targets])
sampler = WeightedRandomSampler(samples_weight, len(samples_weight))


dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,
                                             shuffle=True, num_workers=4, sampler = sampler)
              for x in ['train', 'test']}
dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'test']}


class_names = image_datasets['train'].classes
-------------------------------------------------------------

tensor([ 21, 136, 260,  44, 103,  57, 152,  28,  63])
tensor([0.0476, 0.0074, 0.0038, 0.0227, 0.0097, 0.0175, 0.0066, 0.0357, 0.0159])

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-257-8e92e1fd5cb3> in <module>()
     95 weight = 1. / class_sample_count.float()
     96 print(weight)
---> 97 samples_weight = torch.tensor([weight[t] for t in flat_targets])
     98 sampler = WeightedRandomSampler(samples_weight, len(samples_weight))
     99

<ipython-input-257-8e92e1fd5cb3> in <listcomp>(.0)
     95 weight = 1. / class_sample_count.float()
     96 print(weight)
---> 97 samples_weight = torch.tensor([weight[t] for t in flat_targets])
     98 sampler = WeightedRandomSampler(samples_weight, len(samples_weight))
     99

RuntimeError: tensors used as indices must be long or byte tensors