def __init__(self, args, dropout=0.2):
super().__init__()
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
z_dim = 64
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
z_dim = 640
elif args.model_type == 'AmdimNet':
from feat.networks.amdimnet import AmdimNet
self.encoder = AmdimNet(ndf=args.ndf,
n_rkhs=args.rkhs,
n_depth=args.nd)
z_dim = args.rkhs
else:
raise ValueError('')
self.slf_attn = MultiHeadAttention(args,
args.head,
z_dim,
z_dim,
z_dim,
dropout=dropout)
self.z_dim = z_dim
self.args = args
def __init__(self, args):
super(MatchNet, self).__init__()
self.use_bilstm = args.use_bilstm
self.args = args # information about Shot and Way
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
layer_size = 32
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
layer_size = 320
elif args.model_type == 'AmdimNet':
from feat.networks.amdimnet import AmdimNet
self.encoder = AmdimNet(ndf=args.ndf,
n_rkhs=args.rkhs,
n_depth=args.nd)
layer_size = int(args.rkhs / 2)
else:
raise ValueError('')
if self.use_bilstm:
self.bilstm = BidirectionalLSTM(layer_sizes=[layer_size],
batch_size=args.query * args.way,
vector_dim=layer_size * 2)
def __init__(self, args):
super().__init__()
self.args = args
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
else:
raise ValueError('')
def __init__(self, args):
super().__init__()
self.args = args
if args.model_type == 'ConvNet':
hdim = 64
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
elif args.model_type == 'ResNet':
hdim = 640
from feat.networks.resnet import ResNet as ResNet
self.encoder = ResNet()
else:
raise ValueError('')
self.fc = nn.Linear(hdim, args.num_class)
def __init__(self, args):
super().__init__()
self.args = args
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
elif args.model_type == 'AmdimNet':
from feat.networks.amdimnet import AmdimNet
model = AmdimNet(ndf=args.ndf, n_rkhs=args.rkhs, n_depth=args.nd)
self.encoder = model #torch.nn.DataParallel(model, device_ids=[ torch.device("cuda:0"), torch.device("cuda:1")])
else:
raise ValueError('')
def __init__(self, args, dropout=0.1):
super().__init__()
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
z_dim = 512
self.softmax = nn.Softmax(dim=1)
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
z_dim = 640
else:
raise ValueError('')
self.slf_attn = MultiHeadAttention(1, z_dim, z_dim, z_dim, dropout=dropout)
self.args = args
def __init__(self, args):
super().__init__()
self.args = args
if args.model_type == 'ConvNet':
from feat.networks.convnet import ConvNet
self.encoder = ConvNet()
elif args.model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
elif args.model_type == 'AmdimNet':
from feat.networks.amdimnet import AmdimNet
self.encoder = AmdimNet(ndf=args.ndf,
n_rkhs=args.rkhs,
n_depth=args.nd)
else:
raise ValueError('')
def __init__(self, use_bilstm, model_type, num_shot, num_way, num_query, temperature = 1):
super(MatchNet, self).__init__()
self.use_bilstm = use_bilstm
if model_type == 'ConvNet4':
from feat.networks.convnet import ConvNet4
self.encoder = ConvNet4()
layer_size = 32
elif model_type == 'ResNet':
from feat.networks.resnet import ResNet
self.encoder = ResNet()
layer_size = 320
else:
raise ValueError('')
if self.use_bilstm:
self.bilstm = BidirectionalLSTM(layer_sizes=[layer_size],
batch_size=num_query * num_way,
vector_dim=layer_size * 2)