def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = SSNModel(cfg.fdim, cfg.nspix, cfg.niter).to(device)
optimizer = optim.Adam(model.parameters(), cfg.lr)
augment = augmentation.Compose([
augmentation.RandomHorizontalFlip(),
augmentation.RandomScale(),
augmentation.RandomCrop()
])
train_dataset = bsds.BSDS(cfg.root, geo_transforms=augment)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
test_dataset = bsds.BSDS(cfg.root, split="val")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
max_val_asa = 0
writer = SummaryWriter(log_dir='log', comment='traininglog')
while iterations < cfg.train_iter:
for data in train_loader:
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.color_scale, cfg.pos_scale, device)
meter.add(metric)
state = meter.state(f"[{iterations}/{cfg.train_iter}]")
print(state)
# return {"loss": loss.item(), "reconstruction": recons_loss.item(), "compact": compact_loss.item(),"uniform:":uniform_compactness.item()}
writer.add_scalar("comprehensive/loss", metric["loss"], iterations)
writer.add_scalar("loss/reconstruction_loss",
metric["reconstruction"], iterations)
writer.add_scalar("loss/compact_loss", metric["compact"],
iterations)
if (iterations % cfg.test_interval) == 0:
asa = eval(model, test_loader, cfg.color_scale, cfg.pos_scale,
device)
print(f"validation asa {asa}")
writer.add_scalar("comprehensive/asa", asa, iterations)
if asa > max_val_asa:
max_val_asa = asa
torch.save(
model.state_dict(),
os.path.join(cfg.out_dir, "bset_model_sp_test.pth"))
if iterations == cfg.train_iter:
break
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = MFEAM_SSN(10, 50, backend=soft_slic_pknn).to(device)
print(model)
disc_loss = discriminative_loss(0.1, 0.1)
optimizer = optim.Adam(model.parameters(), cfg.lr)
train_dataset = shapenet.shapenet(cfg.root)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
test_dataset = shapenet.shapenet(cfg.root, split="test")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
max_val_asa = 0
writer = SummaryWriter(log_dir=cfg.out_dir, comment='traininglog')
for epoch_idx in range(cfg.train_epoch):
batch_iterations = 0
for data in train_loader:
batch_iterations += 1
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device, disc_loss)
meter.add(metric)
state = meter.state(
f"[{batch_iterations},{epoch_idx}/{cfg.train_epoch}]")
print(state)
# return {"loss": loss.item(), "reconstruction": recons_loss.item(), "compact": compact_loss.item()}
writer.add_scalar("comprehensive/loss", metric["loss"], iterations)
writer.add_scalar("loss/reconstruction_loss",
metric["reconstruction"], iterations)
writer.add_scalar("loss/compact_loss", metric["compact"],
iterations)
writer.add_scalar("loss/disc_loss", metric["disc"], iterations)
if (iterations % 200) == 0:
(asa, usa) = eval(model, test_loader, cfg.pos_scale, device)
writer.add_scalar("test/asa", asa, iterations)
writer.add_scalar("test/ue", usa, iterations)
if (iterations % 1000) == 0:
strs = "ep_{:}_batch_{:}_iter_{:}_asa_{:.3f}_ue_{:.3f}.pth".format(
epoch_idx, batch_iterations, iterations, asa, usa)
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, strs))
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = LMFEAM_SSN(10, 50, backend=soft_slic_pknn).to(
device) # LMFEAM(10, 50).to(device)
disc_loss = discriminative_loss(0.1, 0.1)
optimizer = optim.Adam(model.parameters(), cfg.lr)
train_dataset = shapenet.shapenet_spix(cfg.root)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
print(train_dataset.__len__())
# test_dataset = shapenet.shapenet(cfg.root, split="test")
# test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
writer = SummaryWriter(log_dir=cfg.out_dir, comment='traininglog')
for epoch_idx in range(cfg.train_epoch):
batch_iterations = 0
for data in train_loader:
iterations += 1
batch_iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device, disc_loss)
meter.add(metric)
state = meter.state(
f"[{batch_iterations},{epoch_idx}/{cfg.train_epoch}]")
print(state)
# return {"loss": loss.item(), "reconstruction": recons_loss.item(), "compact": compact_loss.item()}
writer.add_scalar("comprehensive/loss", metric["loss"], iterations)
writer.add_scalar("loss/spix_loss", metric["spix"], iterations)
writer.add_scalar("loss/reconstruction_loss",
metric["reconstruction"], iterations)
writer.add_scalar("loss/compact_loss", metric["compact"],
iterations)
writer.add_scalar("loss/disc_loss", metric["disc"], iterations)
if (iterations % 500) == 0:
torch.save(
model.state_dict(),
os.path.join(
cfg.out_dir, "model_epoch_" + str(epoch_idx) + "_" +
str(batch_iterations) + '_iter_' + str(iterations) +
".pth"))
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = PointNet_SSN(cfg.fdim,
cfg.nspix,
cfg.niter,
backend=soft_slic_pknn).to(device)
optimizer = optim.Adam(model.parameters(), cfg.lr)
decayer = optim.lr_scheduler.StepLR(optimizer, 1, 0.95)
train_dataset = shapenet.shapenet_inst(cfg.root)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
test_dataset = shapenet.shapenet_inst(cfg.root, split="test")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
writer = SummaryWriter(log_dir=args.out_dir, comment='traininglog')
for epoch_idx in range(cfg.train_epoch):
batch_iterations = 0
for data in train_loader:
batch_iterations += 1
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device)
meter.add(metric)
state = meter.state(
f"[{batch_iterations},{epoch_idx}/{cfg.train_epoch}]")
print(state)
addscaler(metric, writer, iterations)
if (iterations % 200) == 0:
test_res = eval(model, test_loader, cfg.pos_scale, device)
addscaler(test_res, writer, iterations, True)
if (iterations % 1000) == 0:
(asa, usa) = test_res
strs = "ep_{:}_batch_{:}_iter_{:}_asa_{:.3f}_ue_{:.3f}.pth".format(
epoch_idx, batch_iterations, iterations, asa, usa)
torch.save(model.state_dict(),
os.path.join(args.out_dir, strs))
decayer.step()
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = SSNModel(cfg.fdim, cfg.nspix, cfg.niter).to(device)
optimizer = optim.Adam(model.parameters(), cfg.lr)
augment = augmentation.Compose([augmentation.RandomHorizontalFlip(), augmentation.RandomScale(), augmentation.RandomCrop()])
train_dataset = bsds.BSDS(cfg.root, geo_transforms=augment)
train_loader = DataLoader(train_dataset, cfg.batchsize, shuffle=True, drop_last=True, num_workers=cfg.nworkers)
test_dataset = bsds.BSDS(cfg.root, split="val")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
max_val_asa = 0
while iterations < cfg.train_iter:
for data in train_loader:
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness, cfg.color_scale, cfg.pos_scale, device)
meter.add(metric)
state = meter.state(f"[{iterations}/{cfg.train_iter}]")
print(state)
if (iterations % cfg.test_interval) == 0:
asa = eval(model, test_loader, cfg.color_scale, cfg.pos_scale, device)
print(f"validation asa {asa}")
if asa > max_val_asa:
max_val_asa = asa
torch.save(model.state_dict(), os.path.join(cfg.out_dir, "bset_model.pth"))
if iterations == cfg.train_iter:
break
unique_id = str(int(time.time()))
torch.save(model.state_dict(), os.path.join(cfg.out_dir, "model"+unique_id+".pth"))
.
test_dataset = shapenet.shapenet_inst(cfg.root, split="test")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
writer = SummaryWriter(log_dir=cfg.out_dir, comment='traininglog')
for epoch_idx in range(cfg.train_epoch):
batch_iterations = 0
for data in train_loader:
batch_iterations += 1
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device)
meter.add(metric)
state = meter.state(
f"[{batch_iterations},{epoch_idx}/{cfg.train_epoch}]")
print(state)
addscaler(metric, writer, iterations)
if (iterations % 200) == 0:
test_res = eval(model, test_loader, cfg.pos_scale, device)
addscaler(test_res, writer, iterations, True)
if (iterations % 1000) == 0:
(asa, usa) = test_res
strs = "ep_{:}_batch_{:}_iter_{:}_asa_{:.3f}_ue_{:.3f}.pth".format(
epoch_idx, batch_iterations, iterations, asa, usa)
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, strs))
dacayer.step()
unique_id = str(int(time.time()))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = PointNet_SSN(cfg.fdim,
cfg.nspix,
cfg.niter,
backend=soft_slic_pknn).to(device)
optimizer = optim.Adam(model.parameters(), cfg.lr)
train_dataset = shapenet.shapenet(cfg.root)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
test_dataset = shapenet.shapenet(cfg.root, split="test")
test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
max_val_asa = 0
writer = SummaryWriter(log_dir=cfg.out_dir, comment='traininglog')
for epoch_idx in range(cfg.train_epoch):
batch_iterations = 0
for data in train_loader:
batch_iterations += 1
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device)
meter.add(metric)
state = meter.state(
f"[{batch_iterations},{epoch_idx}/{cfg.train_epoch}]")
print(state)
# return {"loss": loss.item(), "reconstruction": recons_loss.item(), "compact": compact_loss.item()}
writer.add_scalar("comprehensive/loss", metric["loss"], iterations)
writer.add_scalar("loss/reconstruction_loss",
metric["reconstruction"], iterations)
writer.add_scalar("loss/compact_loss", metric["compact"],
iterations)
writer.add_scalar("lr", metric["lr"], iterations)
if (iterations % 1000) == 0:
torch.save(
model.state_dict(),
os.path.join(
cfg.out_dir, "model_epoch_" + str(epoch_idx) + "_" +
str(batch_iterations) + '_iter_' + str(iterations) +
".pth"))
# if (iterations % cfg.test_interval) == 0:
# asa = eval(model, test_loader, cfg.pos_scale, device)
# print(f"validation asa {asa}")
# writer.add_scalar("comprehensive/asa", asa, iterations)
# if asa > max_val_asa:
# max_val_asa = asa
# torch.save(model.state_dict(), os.path.join(
# cfg.out_dir, "bset_model_sp_loss.pth"))
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))
def train(cfg):
if torch.cuda.is_available():
device = "cuda"
else:
device = "cpu"
model = MFEAM_SSN(10, 50).to(device)
disc_loss = discriminative_loss(0.1, 0.5)
optimizer = optim.Adam(model.parameters(), cfg.lr)
train_dataset = shapenet_spix(cfg.root)
train_loader = DataLoader(train_dataset,
cfg.batchsize,
shuffle=True,
drop_last=True,
num_workers=cfg.nworkers)
# test_dataset = shapenet.shapenet(cfg.root, split="test")
# test_loader = DataLoader(test_dataset, 1, shuffle=False, drop_last=False)
meter = Meter()
iterations = 0
max_val_asa = 0
writer = SummaryWriter(log_dir='log', comment='traininglog')
while iterations < cfg.train_iter:
for data in train_loader:
iterations += 1
metric = update_param(data, model, optimizer, cfg.compactness,
cfg.pos_scale, device, disc_loss)
meter.add(metric)
state = meter.state(f"[{iterations}/{cfg.train_iter}]")
print(state)
# return {"loss": loss.item(), "reconstruction": recons_loss.item(), "compact": compact_loss.item()}
writer.add_scalar("comprehensive/loss", metric["loss"], iterations)
writer.add_scalar("loss/reconstruction_loss",
metric["reconstruction"], iterations)
writer.add_scalar("loss/compact_loss", metric["compact"],
iterations)
writer.add_scalar("loss/disc_loss", metric["disc"], iterations)
if (iterations % 1000) == 0:
torch.save(
model.state_dict(),
os.path.join(cfg.out_dir,
"model_iter" + str(iterations) + ".pth"))
# if (iterations % cfg.test_interval) == 0:
# asa = eval(model, test_loader, cfg.pos_scale, device)
# print(f"validation asa {asa}")
# writer.add_scalar("comprehensive/asa", asa, iterations)
# if asa > max_val_asa:
# max_val_asa = asa
# torch.save(model.state_dict(), os.path.join(
# cfg.out_dir, "bset_model_sp_loss.pth"))
if iterations == cfg.train_iter:
break
unique_id = str(int(time.time()))
torch.save(model.state_dict(),
os.path.join(cfg.out_dir, "model" + unique_id + ".pth"))