elif args.model == 'lcnn9_mfcc':
model = model_.lcnn_9layers(n_z=args.latent_size,
proj_size=len(train_dataset.speakers_list)
if args.softmax != 'none' else 0,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'lcnn29_mfcc':
model = model_.lcnn_29layers_v2(n_z=args.latent_size,
proj_size=len(train_dataset.speakers_list)
if args.softmax != 'none' else 0,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=len(train_dataset.speakers_list)
if args.softmax != 'none' else 0,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN_multipool':
model = model_.TDNN_multipool(n_z=args.latent_size,
proj_size=len(train_dataset.speakers_list)
if args.softmax != 'none' else 0,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'FTDNN':
model = model_.FTDNN(n_z=args.latent_size,
proj_size=len(train_dataset.speakers_list)
if args.softmax != 'none' else 0,
ncoef=args.ncoef,
sm_type=args.softmax)
proj_size=train_dataset.n_speakers if
args.softmax != 'none' or args.pretrain else 0,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'resnet_2d':
model = model_.ResNet_2d(n_z=args.latent_size,
proj_size=train_dataset.n_speakers
if args.softmax != 'none' or args.pretrain else 0,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=train_dataset.n_speakers
if args.softmax != 'none' or args.pretrain else 0,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN_att':
model = model_.TDNN_att(n_z=args.latent_size,
proj_size=train_dataset.n_speakers
if args.softmax != 'none' or args.pretrain else 0,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN_multihead':
model = model_.TDNN_multihead(
n_z=args.latent_size,
proj_size=train_dataset.n_speakers
if args.softmax != 'none' or args.pretrain else 0,
ncoef=args.ncoef,
ncoef=args.ncoef)
elif args.model == 'resnet_stats':
model = model_.ResNet_stats(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
elif args.model == 'lcnn9_mfcc':
model = model_.lcnn_9layers(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
elif args.model == 'lcnn29_mfcc':
model = model_.lcnn_29layers_v2(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
elif args.model == 'TDNN_multipool':
model = model_.TDNN_multipool(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
elif args.model == 'FTDNN':
model = model_.FTDNN(n_z=args.latent_size,
proj_size=None,
ncoef=args.ncoef)
ckpt = torch.load(args.cp_path, map_location=lambda storage, loc: storage)
model.load_state_dict(ckpt['model_state'], strict=False)
model.eval()
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'lcnn9_mfcc':
model = model_.lcnn_9layers(n_z=args.latent_size,
proj_size=len(list(labels_dict.keys())),
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'lcnn29_mfcc':
model = model_.lcnn_29layers_v2(n_z=args.latent_size,
proj_size=len(list(
labels_dict.keys())),
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=len(list(labels_dict.keys())),
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN_multipool':
model = model_.TDNN_multipool(n_z=args.latent_size,
proj_size=len(list(labels_dict.keys())),
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'FTDNN':
model = model_.FTDNN(n_z=args.latent_size,
proj_size=len(list(labels_dict.keys())),
ncoef=args.ncoef,
sm_type=args.softmax)
if args.cp_path_2 is not None:
model_2 = type(model)(n_z=args.latent_size,
proj_size=len(list(labels_dict.keys())),
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('resnet_small', mu.size(), emb.size(), out.size())
if args.model == 'resnet_2d' or args.model == 'all':
batch = torch.rand(3, 3 if args.delta else 1, 43, 200)
model = model_.ResNet_2d(n_z=args.latent_size,
delta=args.delta,
proj_size=10,
sm_type='softmax')
mu, emb = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('resnet_2d', mu.size(), emb.size(), out.size())
if args.model == 'TDNN' or args.model == 'all':
batch = torch.rand(3, 3 if args.delta else 1, args.ncoef, 200)
model = model_.TDNN(n_z=args.latent_size,
ncoef=args.ncoef,
delta=args.delta,
proj_size=10,
sm_type='softmax')
mu, emb = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('TDNN', mu.size(), emb.size(), out.size())
if args.model == 'TDNN_att' or args.model == 'all':
batch = torch.rand(3, 3 if args.delta else 1, args.ncoef, 200)
model = model_.TDNN_att(n_z=args.latent_size,
ncoef=args.ncoef,
delta=args.delta,
proj_size=10,
sm_type='softmax')
mu, emb = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('TDNN_att', mu.size(), emb.size(), out.size())
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'resnet_small':
model = model_.ResNet_small(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'resnet_2d':
model = model_.ResNet_2d(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'TDNN_att':
model = model_.TDNN_att(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'TDNN_multihead':
model = model_.TDNN_multihead(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
delta=args.delta)
elif args.model == 'TDNN_lstm':
model = model_.TDNN_lstm(n_z=args.latent_size,
proj_size=0,
ncoef=args.ncoef,
if args.model == 'resnet_50' or args.model == 'all':
batch = torch.rand(3, 10000).to(device)
model = model_.ResNet_50(pase_cfg=args.pase_cfg,
pase_cp=args.pase_cp,
n_z=args.latent_size,
ncoef=args.ncoef,
proj_size=10,
sm_type='softmax').to(device)
mu = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('resnet_50', mu.size(), out.size())
if args.model == 'TDNN' or args.model == 'all':
batch = torch.rand(3, 10000).to(device)
model = model_.TDNN(pase_cfg=args.pase_cfg,
pase_cp=args.pase_cp,
n_z=args.latent_size,
ncoef=args.ncoef,
proj_size=10,
sm_type='softmax').to(device)
mu = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('TDNN', mu.size(), out.size())
if args.model == 'TDNN_mfcc' or args.model == 'all':
batch = torch.rand(3, args.ncoef, 200).to(device)
model = model_.TDNN_mfcc(n_z=args.latent_size,
ncoef=args.ncoef,
proj_size=10,
sm_type='softmax').to(device)
mu = model.forward(batch)
out = model.out_proj(mu, torch.ones(mu.size(0)))
print('TDNN_mfcc', mu.size(), out.size())
if args.model == 'MLP' or args.model == 'all':
nh=ckpt['n_hidden'],
n_h=ckpt['hidden_size'],
proj_size=ckpt['r_proj_size'],
ncoef=ckpt['ncoef'],
ndiscriminators=ckpt['ndiscriminators'])
elif args.model == 'resnet_large':
model = model_.ResNet_large(n_z=ckpt['latent_size'],
nh=ckpt['n_hidden'],
n_h=ckpt['hidden_size'],
proj_size=ckpt['r_proj_size'],
ncoef=ckpt['ncoef'],
ndiscriminators=ckpt['ndiscriminators'])
elif args.model == 'TDNN':
model = model_.TDNN(n_z=ckpt['latent_size'],
nh=ckpt['n_hidden'],
n_h=ckpt['hidden_size'],
proj_size=ckpt['r_proj_size'],
ncoef=ckpt['ncoef'],
ndiscriminators=ckpt['ndiscriminators'])
try:
model.load_state_dict(ckpt['model_state'], strict=True)
except RuntimeError as err:
print("Runtime Error: {0}".format(err))
except:
print("Unexpected error:", sys.exc_info()[0])
raise
model.eval()
if args.cuda:
model = model.to(device)
elif args.model == 'resnet_large':
model = model_.ResNet_large(n_z=args.latent_size,
nh=args.n_hidden,
n_h=args.hidden_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
dropout_prob=args.dropout_prob,
sm_type=args.softmax,
ndiscriminators=args.ndiscriminators,
r_proj_size=args.rproj_size)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
nh=args.n_hidden,
n_h=args.hidden_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
dropout_prob=args.dropout_prob,
sm_type=args.softmax,
ndiscriminators=args.ndiscriminators,
r_proj_size=args.rproj_size)
if args.cuda:
device = get_freer_gpu()
else:
device = None
if args.cuda:
model = model.cuda(device)
optimizer = TransformerOptimizer(optim.SGD(model.parameters(),
lr=args.lr,
model = model_.ResNet_stats(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('resnet_stats', mu.size())
if args.model == 'lcnn9_mfcc' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 400)
model = model_.lcnn_9layers(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('lcnn9_mfcc', mu.size())
if args.model == 'lcnn29_mfcc' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 400)
model = model_.lcnn_29layers_v2(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('lcnn29_mfcc', mu.size())
if args.model == 'TDNN' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 400)
model = model_.TDNN(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('TDNN', mu.size())
if args.model == 'TDNN_multipool' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 400)
model = model_.TDNN_multipool(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('TDNN_multipool', mu.size())
if args.model == 'FTDNN' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 400)
model = model_.FTDNN(n_z=args.latent_size, ncoef=args.ncoef)
mu = model.forward(batch)
print('FTDNN', mu.size())
if args.softmax:
batch = torch.rand(3, mu.size(0))
ncoef=args.ncoef,
ndiscriminators=args.ndiscriminators,
r_proj_size=args.rproj_size)
print('resnet_large')
mu, emb = model.forward(batch)
print(mu.size())
emb = torch.cat([emb, emb], 1)
print(emb.size())
pred = model.forward_bin(emb)
print(pred)
scores_p = model.forward_bin(emb)
print(scores_p)
if args.model == 'TDNN' or args.model == 'all':
batch = torch.rand(3, 1, args.ncoef, 200)
model = model_.TDNN(n_z=args.latent_size,
nh=args.n_hidden,
n_h=args.hidden_size,
proj_size=100,
ncoef=args.ncoef,
ndiscriminators=args.ndiscriminators,
r_proj_size=args.rproj_size)
print('TDNN')
mu, emb = model.forward(batch)
print(mu.size())
emb = torch.cat([emb, emb], 1)
print(emb.size())
pred = model.forward_bin(emb)
print(pred)
scores_p = model.forward_bin(emb)
print(scores_p)
elif args.model == 'resnet_small':
model = model_.ResNet_small(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'resnet_2d':
model = model_.ResNet_2d(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN_logpool':
model = model_.TDNN_logpool(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN_att':
model = model_.TDNN_att(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax,
delta=args.delta)
elif args.model == 'TDNN_multihead':
if args.cp_path is None:
raise ValueError('There is no checkpoint/model path. Use arg --cp-path to indicate the path!')
print('Cuda Mode is: {}'.format(args.cuda))
if args.cuda:
device = get_freer_gpu()
if args.model == 'resnet_18':
model = model_.ResNet_18(pase_cfg=args.pase_cfg, n_z=args.latent_size, proj_size=None, ncoef=args.ncoef)
elif args.model == 'resnet_34':
model = model_.ResNet_34(pase_cfg=args.pase_cfg, n_z=args.latent_size, proj_size=None, ncoef=args.ncoef)
elif args.model == 'resnet_50':
model = model_.ResNet_50(pase_cfg=args.pase_cfg, n_z=args.latent_size, proj_size=None, ncoef=args.ncoef)
elif args.model == 'TDNN':
model = model_.TDNN(pase_cfg=args.pase_cfg, n_z=args.latent_size, proj_size=None, ncoef=args.ncoef)
ckpt = torch.load(args.cp_path, map_location = lambda storage, loc: storage)
model.load_state_dict(ckpt['model_state'], strict=False)
model.eval()
if args.cuda:
model = model.to(device)
enroll_utt_data = read_utt2rec(args.enroll_data+'wav.scp', args.m4a)
test_utt_data = read_utt2rec(args.test_data+'wav.scp', args.m4a)
utterances_enroll, utterances_test, labels = read_trials(args.trials_path)
print('\nAll data ready. Start of scoring')
def train(lr, l2, max_gnorm, momentum, margin, lambda_, swap, latent_size, n_frames, model, ncoef, epochs, batch_size, valid_batch_size, n_workers, cuda, train_hdf_file, valid_hdf_file, cp_path, softmax, delta, logdir):
if cuda:
device=get_freer_gpu()
if args.model == 'resnet_qrnn':
import cupy
cupy.cuda.Device(int(str(device).split(':')[-1])).use()
cp_name = get_file_name(cp_path)
if args.logdir:
from torch.utils.tensorboard import SummaryWriter
writer = SummaryWriter(log_dir=logdir+cp_name, comment=args.model, purge_step=True)
else:
writer = None
train_dataset = Loader(hdf5_name = train_hdf_file, max_nb_frames = int(n_frames), delta = delta)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=n.workers, worker_init_fn=set_np_randomseed)
valid_dataset = Loader_valid(hdf5_name = valid_hdf_file, max_nb_frames = int(n_frames), delta = delta)
valid_loader = torch.utils.data.DataLoader(valid_dataset, batch_size=valid_batch_size, shuffle=True, num_workers=n_workers, worker_init_fn=set_np_randomseed)
if model == 'resnet_mfcc':
model=model_.ResNet_mfcc(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=ncoef, sm_type=softmax, delta=delta)
elif model == 'resnet_34':
model=model_.ResNet_34(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=ncoef, sm_type=softmax, delta=delta)
elif model == 'resnet_lstm':
model=model_.ResNet_lstm(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=ncoef, sm_type=softmax, delta=delta)
elif model == 'resnet_qrnn':
model=model_.ResNet_qrnn(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=ncoef, sm_type=softmax, delta=delta)
elif model == 'resnet_stats':
model=model_.ResNet_stats(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=ncoef, sm_type=softmax, delta=delta)
elif args.model == 'resnet_large':
model = model_.ResNet_large(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'resnet_small':
model = model_.ResNet_small(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'resnet_2d':
model = model_.ResNet_2d(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN':
model = model_.TDNN(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_att':
model = model_.TDNN_att(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_multihead':
model = model_.TDNN_multihead(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_lstm':
model = model_.TDNN_lstm(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_aspp':
model = model_.TDNN_aspp(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_mod':
model = model_.TDNN_mod(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'TDNN_multipool':
model = model_.TDNN_multipool(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
elif args.model == 'transformer':
model = model_.transformer_enc(n_z=int(latent_size), proj_size=train_dataset.n_speakers, ncoef=args.ncoef, sm_type=softmax, delta=delta)
if cuda:
model=model.to(device)
else:
device=None
optimizer=optim.SGD(model.parameters(), lr=lr, momentum=momentum, weight_decay=l2)
trainer=TrainLoop(model, optimizer, train_loader, valid_loader, max_gnorm=max_gnorm, margin=margin, lambda_=lambda_, verbose=-1, device=device, cp_name=cp_name, save_cp=True, checkpoint_path=cp_path, swap=swap, softmax=True, pretrain=False, mining=True, cuda=cuda, logger=writer)
return trainer.train(n_epochs=epochs)
pase_cp=args.pase_cp,
n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'resnet_50':
model = model_.ResNet_50(pase_cfg=args.pase_cfg,
pase_cp=args.pase_cp,
n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN':
model = model_.TDNN(pase_cfg=args.pase_cfg,
pase_cp=args.pase_cp,
n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'TDNN_mfcc':
model = model_.TDNN_mfcc(n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'MLP':
model = model_.MLP(pase_cfg=args.pase_cfg,
pase_cp=args.pase_cp,
n_z=args.latent_size,
proj_size=train_dataset.n_speakers,
ncoef=args.ncoef,
sm_type=args.softmax)
elif args.model == 'global_MLP':
