in eft/datasets/mixed_dataset.py [0:0]
def __init__(self, options, **kwargs):
# self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet', 'coco', 'mpi-inf-3dhp']
# self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
# self.dataset_list = ['lsp-orig', 'mpii', 'lspet', 'coco', 'mpi-inf-3dhp']
# self.dataset_dict = {'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
# self.dataset_list = ['lsp-orig', 'mpii', 'lspet', 'coco']
# self.dataset_dict = {'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4}
# self.dataset_list = ['coco', 'mpi-inf-3dhp']
# self.dataset_dict = {'coco': 4, 'mpi-inf-3dhp': 5}
# self.dataset_list = ['mpi-inf-3dhp']
# self.dataset_dict = {'mpi-inf-3dhp': 5}
# self.dataset_list = [ 'mpii', 'coco']
# self.dataset_dict = { 'mpii': 2, 'coco': 4}
self.partition =[]
length_itw =0
if options.db_set == 'coco':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco']
self.dataset_dict = { 'coco': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco-val':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco-val']
self.dataset_dict = { 'coco-val': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocofoot':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocofoot']
self.dataset_dict = { 'cocofoot': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocofoot3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocofoot3d']
self.dataset_dict = { 'cocofoot3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'ochuman':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'ochuman']
self.dataset_dict = { 'ochuman': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'ochuman3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'ochuman3d']
self.dataset_dict = { 'ochuman3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall']
self.dataset_dict = { 'cocoall': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d']
self.dataset_dict = { 'cocoall3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoplus3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoplus3d']
self.dataset_dict = { 'cocoplus3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'posetrack':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'posetrack']
self.dataset_dict = { 'posetrack': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'posetrack3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'posetrack3d']
self.dataset_dict = { 'posetrack3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocopart_posetrack3d':
self.dataset_list = ['coco3d', 'posetrack3d']
self.dataset_dict = {'coco3d': 0, 'posetrack3d': 1}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall_posetrack3d':
self.dataset_list = ['cocoall3d', 'posetrack3d']
self.dataset_dict = {'cocoall3d': 0, 'posetrack3d': 1}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'multicrop_cocoall_posetrack3d':
self.dataset_list = ['multicrop_cocoall3d', 'multicrop_posetrack3d']
self.dataset_dict = {'multicrop_cocoall3d': 0, 'multicrop_posetrack3d': 1}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall_posetrack3d_lspetatm':
self.dataset_list = ['cocoall3d', 'posetrack3d', 'lspet3d-amt']
self.dataset_dict = {'cocoall3d': 0, 'posetrack3d': 1, 'lspet3d-amt':2}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw
,len(self.datasets[2])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall_posetrack3d_lspetatm-train':
self.dataset_list = ['cocoall3d', 'posetrack3d', 'lspet3d-amt-train']
self.dataset_dict = {'cocoall3d': 0, 'posetrack3d': 1, 'lspet3d-amt-train':2}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw
,len(self.datasets[2])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall_mpii':
self.dataset_list = ['cocoall3d', 'mpii3d']
self.dataset_dict = {'cocoall3d': 0, 'mpii3d': 1}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall_mpii_posetrack3d':
self.dataset_list = ['cocoall3d', 'mpii3d', 'posetrack3d']
self.dataset_dict = {'cocoall3d': 0, 'mpii3d': 1, 'posetrack3d': 2}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw,
len(self.datasets[2])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == '3dpw_test':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test']
self.dataset_dict = { '3dpw_test': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_test_crop':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test_crop']
self.dataset_dict = { '3dpw_test_crop': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_test_multilevel_0':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test_multilevel_0']
self.dataset_dict = { '3dpw_test_multilevel_0': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_test_multilevel_1':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test_multilevel_1']
self.dataset_dict = { '3dpw_test_multilevel_1': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_test_multilevel_4':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test_multilevel_1']
self.dataset_dict = { '3dpw_test_multilevel_1': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_test_multilevel_7':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_test_multilevel_7']
self.dataset_dict = { '3dpw_test_multilevel_7': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == '3dpw_train':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ '3dpw_train']
self.dataset_dict = { '3dpw_train': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'h36m_p2_test':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'h36m-p2']
self.dataset_dict = { 'h36m-p2': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'h36m_p2_test':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'h36m-p2']
self.dataset_dict = { 'h36m-p2': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'pennaction':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'pennaction']
self.dataset_dict = { 'pennaction': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'pennaction_all':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'lsp-orig', 'mpii', 'lspet', 'coco', 'pennaction' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'lsp-orig': 2, 'mpii': 3, 'lspet': 4, 'coco': 5, 'pennaction': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw,
.6*len(self.datasets[6])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'panoptic':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'panoptic']
self.dataset_dict = { 'panoptic': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'panoptic3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'panoptic3d']
self.dataset_dict = { 'panoptic3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco3d_panoptic3d': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d', 'panoptic3d']
self.dataset_dict = { 'coco3d': 0, 'panoptic3d': 1}
self.partition = [0.4, 0.6]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_panoptic3d': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'panoptic3d']
self.dataset_dict = { 'cocoall3d': 0, 'panoptic3d': 1}
self.partition = [0.4, 0.6]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_pan3d_h36m': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'panoptic3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'panoptic3d': 1, 'h36m': 2}
self.partition = [0.4, 0.2, 0.4]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_pan3d_h36m_ochuman3d': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'panoptic3d', 'h36m','cocoall3d', 'ochuman3d']
self.dataset_dict = { 'panoptic3d': 0, 'h36m': 1, 'cocoall3d': 2, 'ochuman3d': 3}
self.partition = [0.4, 0.1, 0.4, 0.1]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
# self.partition = [.1, 0.3,
# .6*len(self.datasets[2])/length_itw,
# .6*len(self.datasets[3])/length_itw]
self.partition = [.1, 0.3, 0.4, 0.2]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d_pan3d_h36m': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d', 'panoptic3d', 'h36m']
self.dataset_dict = { 'coco3d': 0, 'panoptic3d': 1, 'h36m': 2}
self.partition = [0.4, 0.2, 0.4]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'panoptichand':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'panoptichand']
self.dataset_dict = { 'panoptichand': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'panoptic_haggling_test':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'panoptic_haggling_test']
self.dataset_dict = { 'panoptic_haggling_test': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'panoptic_all':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'panoptic', 'lsp-orig', 'mpii', 'lspet', 'coco' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'panoptic':2, 'lsp-orig': 3, 'mpii': 4, 'lspet': 5, 'coco': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[3:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.2, 0.1, 0.1,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw,
.6*len(self.datasets[6])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'h36m':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'h36m']
self.dataset_dict = { 'h36m': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'lsp':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'lsp-orig']
self.dataset_dict = { 'lsp-orig': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'lspet':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'lspet']
self.dataset_dict = { 'lspet': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'lspet3d-amt':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'lspet3d-amt']
self.dataset_dict = { 'lspet3d-amt': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'lspet3d-amt-train':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'lspet3d-amt-train']
self.dataset_dict = { 'lspet3d-amt-train': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'minf':
self.dataset_list = [ 'mpi-inf-3dhp']
self.dataset_dict = { 'mpi-inf-3dhp': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'mpii':
self.dataset_list = [ 'mpii']
self.dataset_dict = { 'mpii': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'ori_all_noinf':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet', 'coco']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.35, .65*len(self.datasets[1])/length_itw,
.65*len(self.datasets[2])/length_itw,
.65*len(self.datasets[3])/length_itw,
.65*len(self.datasets[4])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'ori_all_noh36m': #original all
self.dataset_list = ['lsp-orig', 'mpii', 'lspet', 'coco', 'mpi-inf-3dhp']
self.dataset_dict = {'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.8*len(self.datasets[0])/length_itw,
.8*len(self.datasets[1])/length_itw,
.8*len(self.datasets[2])/length_itw,
.8*len(self.datasets[3])/length_itw,
0.2]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'ori_coco_h36m_inf':
self.dataset_list = ['h36m','coco', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0,'coco': 1, 'mpi-inf-3dhp': 2}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:2]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'ori_all': #original all including h36m
self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet', 'coco', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d']
self.dataset_dict = { 'coco3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
# length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco3d_amt': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d_amt']
self.dataset_dict = { 'coco3d_amt': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
#This dataset has multiple crop information (8 levels, 0 for head 8 for whole)
elif options.db_set == 'multicrop_coco3dpart':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'multicrop_coco3d']
self.dataset_dict = { 'multicrop_coco3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
#This dataset has multiple crop information (8 levels, 0 for head 8 for whole)
elif options.db_set == 'multicrop_cocoall3d':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'multicrop_cocoall3d']
self.dataset_dict = { 'multicrop_cocoall3d': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco2014_train_6kp_semmap': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco2014_train_6kp_semmap']
self.dataset_dict = { 'coco2014_train_6kp_semmap': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco2017_whole_train_6kp': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco2017_whole_train_6kp']
self.dataset_dict = { 'coco2017_whole_train_6kp': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco2017_whole_train_12kp': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco2017_whole_train_12kp']
self.dataset_dict = { 'coco2017_whole_train_12kp': 0}
self.partition = [1.0]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco3d_cocoplus3d': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d', 'cocoplus3d']
self.dataset_dict = { 'coco3d': 0, 'cocoplus3d': 1}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d_h36m': #original all including h36m
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d', 'h36m']
self.dataset_dict = { 'coco3d': 0, 'h36m': 1}
self.partition = [0.4, 0.6]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
#Default
elif options.db_set == 'cocoall3d_h36m':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.4, 0.6]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_06':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.4, 0.6]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_05':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.5, 0.5]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_04':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.6, 0.4]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_03':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.7, 0.3]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_02':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.8, 0.2]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'cocoall3d_h36m_01':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1}
self.partition = [0.9, 0.1]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
elif options.db_set == 'coco3d_all': #original all including h36m
self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mpii3d':
self.dataset_list = ['mpii3d']
self.dataset_dict = {'mpii3d': 0}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [1.0]
elif options.db_set == 'mpii3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii3d', 'lspet', 'coco', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii3d': 2, 'lspet': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d_mpii3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii3d', 'lspet', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii3d': 2, 'lspet': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'lspet3d':
self.dataset_list = ['lspet3d']
self.dataset_dict = {'lspet3d': 0}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [1.0]
elif options.db_set == 'lspet3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet3d', 'coco', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet3d': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mpii3d_lspet3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii3d', 'lspet3d', 'coco', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mlc3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii3d', 'lspet3d', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mllc3d_all':
self.dataset_list = ['h36m', 'lsp-orig3d', 'mpii3d', 'lspet3d', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig3d': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
################################################################################################################
elif options.db_set == 'eft_spindb_all':
self.dataset_list = ['h36m', 'mpii3d', 'lspet3d', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'mpii3d': 1, 'lspet3d': 2, 'coco3d': 3, 'mpi-inf-3dhp': 4}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3,
.6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
0.1]
print("\nsampling rate: {}".format(self.partition))
elif options.db_set == 'eft_spindb_all_pan3d':
self.dataset_list = ['h36m', 'mpii3d', 'lspet3d', 'coco3d', 'mpi-inf-3dhp', 'panoptic3d']
self.dataset_dict = {'h36m': 0, 'mpii3d': 1, 'lspet3d': 2, 'coco3d': 3, 'mpi-inf-3dhp': 4, 'panoptic3d': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-2]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3,
.55*len(self.datasets[1])/length_itw,
.55*len(self.datasets[2])/length_itw,
.55*len(self.datasets[3])/length_itw,
0.1, 0.05]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'eft_spindb_all_cocoall':
self.dataset_list = ['h36m', 'mpii3d', 'lspet3d', 'cocoall3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'mpii3d': 1, 'lspet3d': 2, 'cocoall3d': 3, 'mpi-inf-3dhp': 4}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3,
.6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'eft_spindb_all_pan3d_cocoall':
self.dataset_list = ['h36m', 'mpii3d', 'lspet3d', 'cocoall3d', 'mpi-inf-3dhp', 'panoptic3d']
self.dataset_dict = {'h36m': 0, 'mpii3d': 1, 'lspet3d': 2, 'cocoall3d': 3, 'mpi-inf-3dhp': 4, 'panoptic3d': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-2]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3,
.55*len(self.datasets[1])/length_itw,
.55*len(self.datasets[2])/length_itw,
.55*len(self.datasets[3])/length_itw,
0.1, 0.05]
print("sampling rate: {}".format(self.partition))
################################################################################################################
elif options.db_set == 'panoptic_ours3d_all':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'panoptic3d', 'lsp-orig', 'mpii3d', 'lspet3d', 'coco3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'panoptic3d':2, 'lsp-orig': 3, 'mpii3d': 4, 'lspet3d': 5, 'coco3d': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[3:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1, 0.1,
.5*len(self.datasets[3])/length_itw,
.5*len(self.datasets[4])/length_itw,
.5*len(self.datasets[5])/length_itw,
.5*len(self.datasets[6])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mc3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii3d', 'lspet', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii3d': 2, 'lspet': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'lc3d_all':
self.dataset_list = ['h36m', 'lsp-orig', 'mpii', 'lspet3d', 'coco3d', 'mpi-inf-3dhp']
self.dataset_dict = {'h36m': 0, 'lsp-orig': 1, 'mpii': 2, 'lspet3d': 3, 'coco3d': 4, 'mpi-inf-3dhp': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[1:-1]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, .6*len(self.datasets[1])/length_itw,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
0.1]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'lc3d_all_cocoplus':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'lspet3d', 'coco3d', 'cocoplus3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'lspet3d': 2, 'coco3d': 3, 'cocoplus3d': 4}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'lc3d_all_posetrack':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'lspet3d', 'coco3d', 'posetrack3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'lspet3d': 2, 'coco3d': 3, 'posetrack3d': 4}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'lc3d_all_cp_pt':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'lspet3d', 'coco3d', 'cocoplus3d', 'posetrack3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'lspet3d': 2, 'coco3d': 3, 'cocoplus3d': 4, 'posetrack3d': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw,
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mlc3d_all_cocoplus':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'mpii3d', 'lspet3d', 'coco3d', 'cocoplus3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco3d': 4, 'cocoplus3d': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mlc3d_all_posetrack':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'mpii3d', 'lspet3d', 'coco3d', 'posetrack3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco3d': 4, 'posetrack3d': 5}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'panoptic_ours3d_all_test':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'panoptic', 'lsp-orig', 'mpii3d', 'lspet3d', 'coco3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'panoptic':2, 'lsp-orig': 3, 'mpii3d': 4, 'lspet3d': 5, 'coco3d': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[3:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.2, 0.0, 0.3,
.5*len(self.datasets[3])/length_itw,
.5*len(self.datasets[4])/length_itw,
.5*len(self.datasets[5])/length_itw,
.5*len(self.datasets[6])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'panoptic_ours3d_all_test2':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'panoptic', 'lsp-orig', 'mpii3d', 'lspet3d', 'coco3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'panoptic':2, 'lsp-orig': 3, 'mpii3d': 4, 'lspet3d': 5, 'coco3d': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[3:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.1, 0.0, 0.4,
.5*len(self.datasets[3])/length_itw,
.5*len(self.datasets[4])/length_itw,
.5*len(self.datasets[5])/length_itw,
.5*len(self.datasets[6])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'mlc3d_all_cp_ps':
self.dataset_list = ['h36m', 'mpi-inf-3dhp', 'mpii3d', 'lspet3d', 'coco3d', 'cocoplus3d', 'posetrack3d' ]
self.dataset_dict = {'h36m': 0, 'mpi-inf-3dhp': 1, 'mpii3d': 2, 'lspet3d': 3, 'coco3d': 4, 'cocoplus3d': 5, 'posetrack3d': 6}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets[2:]])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [.3, 0.1,
.6*len(self.datasets[2])/length_itw,
.6*len(self.datasets[3])/length_itw,
.6*len(self.datasets[4])/length_itw,
.6*len(self.datasets[5])/length_itw,
.6*len(self.datasets[6])/length_itw
]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d_plus3d_mpii3d':
self.dataset_list = ['coco3d','cocoplus3d', 'mpii3d']
self.dataset_dict = {'coco3d': 0, 'cocoplus3d': 1, 'mpii3d': 2}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw,
len(self.datasets[2])/length_itw]
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'coco3d_plus3d_mpii3d_pt3d':
self.dataset_list = ['coco3d','cocoplus3d', 'mpii3d', 'posetrack3d']
self.dataset_dict = {'coco3d': 0, 'cocoplus3d': 1, 'mpii3d': 2, 'posetrack3d': 3}
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
length_itw = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
self.partition = [len(self.datasets[0])/length_itw,
len(self.datasets[1])/length_itw,
len(self.datasets[2])/length_itw,
len(self.datasets[3])/length_itw,]
print("sampling rate: {}".format(self.partition))
################ w/ MPI_INF ################
elif options.db_set == 'coco3d_h36m_inf':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'coco3d', 'h36m', 'mpi-inf-3dhp']
self.dataset_dict = { 'coco3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2}
self.partition = [0.3, 0.5, 0.2]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2}
self.partition = [0.4, 0.5, 0.1]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_3dpw':
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', '3dpw_train']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, '3dpw_train':3}
# self.partition = [0.4, 0.4, 0.1, 0.1]
self.partition = [0.3, 0.4, 0.1, 0.2]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_posetrack3d': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', 'posetrack3d']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'posetrack3d':3}
self.partition = [0.4, 0.4, 0.1, 0.1]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_3dpw_posetrack3d': #All available including 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', '3dpw_train', 'posetrack3d']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, '3dpw_train':3, 'posetrack3d':4}
self.partition = [0.3, 0.4, 0.1, 0.1, 0.1]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_posetrack3d_lspamttrain': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', 'posetrack3d','lspet3d-amt-train']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'posetrack3d':3, 'lspet3d-amt-train':4}
self.partition = [0.4, 0.35, 0.1, 0.1, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'multicrop_cocoall3d_h36m_inf_posetrack3d_lspamttrain': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'multicrop_cocoall3d', 'h36m', 'mpi-inf-3dhp', 'multicrop_posetrack3d','lspet3d-amt-train']
self.dataset_dict = { 'multicrop_cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'multicrop_posetrack3d':3, 'lspet3d-amt-train':4}
self.partition = [0.4, 0.35, 0.1, 0.1, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_3dpw_posetrack3d_lspamttrain': #All available including 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', '3dpw_train', 'posetrack3d', 'lspet3d-amt-train']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, '3dpw_train':3, 'posetrack3d':4, 'lspet3d-amt-train':5}
self.partition = [0.3, 0.35, 0.1, 0.1, 0.1, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_posetrack3d_lspamttrain_ochumantrain': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', 'posetrack3d','lspet3d-amt-train','ochuman3d']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'posetrack3d':3, 'lspet3d-amt-train':4, 'ochuman3d':5}
self.partition = [0.4, 0.3, 0.1, 0.1, 0.05, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'multicrop_cocoall3d_h36m_inf_posetrack3d_lspamttrain_ochumantrain': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'multicrop_cocoall3d', 'h36m', 'mpi-inf-3dhp', 'multicrop_posetrack3d','lspet3d-amt-train','ochuman3d']
self.dataset_dict = { 'multicrop_cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'multicrop_posetrack3d':3, 'lspet3d-amt-train':4, 'ochuman3d':5}
self.partition = [0.4, 0.3, 0.1, 0.1, 0.05, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_3dpw_posetrack3d_lspamttrain_ochumantrain': #All available including 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', '3dpw_train', 'posetrack3d', 'lspet3d-amt-train','ochuman3d']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, '3dpw_train':3, 'posetrack3d':4, 'lspet3d-amt-train':5, 'ochuman3d':6}
self.partition = [0.3, 0.3, 0.1, 0.1, 0.1, 0.05, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_posetrack3d_lspamtall': #All available without 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', 'posetrack3d','lspet3d-amt']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, 'posetrack3d':3, 'lspet3d-amt':4}
self.partition = [0.4, 0.35, 0.1, 0.1, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
elif options.db_set == 'cocoall3d_h36m_inf_3dpw_posetrack3d_lspamtall': #All available including 3dpw
print(">>> Selected DBSet: {}".format(options.db_set))
self.dataset_list = [ 'cocoall3d', 'h36m', 'mpi-inf-3dhp', '3dpw_train', 'posetrack3d', 'lspet3d-amt']
self.dataset_dict = { 'cocoall3d': 0, 'h36m': 1, 'mpi-inf-3dhp':2, '3dpw_train':3, 'posetrack3d':4, 'lspet3d-amt':5}
self.partition = [0.3, 0.35, 0.1, 0.1, 0.1, 0.05]
self.datasets = [BaseDataset(options, ds, **kwargs) for ds in self.dataset_list]
total_length = sum([len(ds) for ds in self.datasets])
self.length = max([len(ds) for ds in self.datasets])
print("sampling rate: {}".format(self.partition))
else:
assert False
assert len(self.partition) == len(self.dataset_list)
print(">>> Total DB num: {} | total in-the-wild DB num: {}".format(total_length, length_itw))
"""
Data distribution inside each batch:
30% H36M - 60% ITW - 10% MPI-INF
"""
# self.partition = [0.1, 0.2, 0.2, 0.5]
# self.partition = [.8*len(self.datasets[0])/length_itw,
# .8*len(self.datasets[1])/length_itw,
# .8*len(self.datasets[2])/length_itw,
# .8*len(self.datasets[3])/length_itw,
# 0.2]
self.partition = np.array(self.partition).cumsum()
#Save Dataset information as log
with open(join(options.log_dir, "dbinfo.json"), "w") as f:
json.dump(config.DATASET_FILES[1], f, indent=4) #config.DATASET_FILES[1] has training info