Update xmisc
This commit is contained in:
D-X-Y
@@ -58,6 +58,7 @@ def main(args):
|
||||
pin_memory=True,
|
||||
drop_last=False,
|
||||
)
|
||||
iters_per_epoch = len(train_data) // args.batch_size
|
||||
|
||||
logger.log("The training loader: {:}".format(train_loader))
|
||||
logger.log("The validation loader: {:}".format(valid_loader))
|
||||
@@ -67,159 +68,44 @@ def main(args):
|
||||
lr=args.lr,
|
||||
weight_decay=args.weight_decay,
|
||||
)
|
||||
loss = xmisc.nested_call_by_yaml(args.loss_config)
|
||||
objective = xmisc.nested_call_by_yaml(args.loss_config)
|
||||
|
||||
logger.log("The optimizer is:\n{:}".format(optimizer))
|
||||
logger.log("The loss is {:}".format(loss))
|
||||
logger.log("The objective is {:}".format(objective))
|
||||
logger.log("The iters_per_epoch={:}".format(iters_per_epoch))
|
||||
|
||||
model, loss = torch.nn.DataParallel(model).cuda(), loss.cuda()
|
||||
model, objective = torch.nn.DataParallel(model).cuda(), objective.cuda()
|
||||
scheduler = xmisc.LRMultiplier(
|
||||
optimizer, xmisc.get_scheduler(args.scheduler, args.lr), args.steps
|
||||
)
|
||||
|
||||
import pdb
|
||||
|
||||
pdb.set_trace()
|
||||
|
||||
train_func, valid_func = get_procedures(args.procedure)
|
||||
|
||||
total_epoch = optim_config.epochs + optim_config.warmup
|
||||
# Main Training and Evaluation Loop
|
||||
start_time = time.time()
|
||||
epoch_time = AverageMeter()
|
||||
for epoch in range(start_epoch, total_epoch):
|
||||
scheduler.update(epoch, 0.0)
|
||||
start_time, iter_time = time.time(), xmisc.AverageMeter()
|
||||
for xiter, data in enumerate(train_loader):
|
||||
need_time = "Time Left: {:}".format(
|
||||
convert_secs2time(epoch_time.avg * (total_epoch - epoch), True)
|
||||
)
|
||||
epoch_str = "epoch={:03d}/{:03d}".format(epoch, total_epoch)
|
||||
LRs = scheduler.get_lr()
|
||||
find_best = False
|
||||
# set-up drop-out ratio
|
||||
if hasattr(base_model, "update_drop_path"):
|
||||
base_model.update_drop_path(
|
||||
model_config.drop_path_prob * epoch / total_epoch
|
||||
)
|
||||
logger.log(
|
||||
"\n***{:s}*** start {:s} {:s}, LR=[{:.6f} ~ {:.6f}], scheduler={:}".format(
|
||||
time_string(), epoch_str, need_time, min(LRs), max(LRs), scheduler
|
||||
xmisc.time_utils.convert_secs2time(
|
||||
iter_time.avg * (len(train_loader) - xiter), True
|
||||
)
|
||||
)
|
||||
iter_str = "{:6d}/{:06d}".format(xiter, len(train_loader))
|
||||
|
||||
# train for one epoch
|
||||
train_loss, train_acc1, train_acc5 = train_func(
|
||||
train_loader,
|
||||
network,
|
||||
criterion,
|
||||
scheduler,
|
||||
optimizer,
|
||||
optim_config,
|
||||
epoch_str,
|
||||
args.print_freq,
|
||||
logger,
|
||||
)
|
||||
# log the results
|
||||
logger.log(
|
||||
"***{:s}*** TRAIN [{:}] loss = {:.6f}, accuracy-1 = {:.2f}, accuracy-5 = {:.2f}".format(
|
||||
time_string(), epoch_str, train_loss, train_acc1, train_acc5
|
||||
)
|
||||
)
|
||||
inputs, targets = data
|
||||
targets = targets.cuda(non_blocking=True)
|
||||
model.train()
|
||||
|
||||
# evaluate the performance
|
||||
if (epoch % args.eval_frequency == 0) or (epoch + 1 == total_epoch):
|
||||
logger.log("-" * 150)
|
||||
valid_loss, valid_acc1, valid_acc5 = valid_func(
|
||||
valid_loader,
|
||||
network,
|
||||
criterion,
|
||||
optim_config,
|
||||
epoch_str,
|
||||
args.print_freq_eval,
|
||||
logger,
|
||||
)
|
||||
valid_accuracies[epoch] = valid_acc1
|
||||
logger.log(
|
||||
"***{:s}*** VALID [{:}] loss = {:.6f}, accuracy@1 = {:.2f}, accuracy@5 = {:.2f} | Best-Valid-Acc@1={:.2f}, Error@1={:.2f}".format(
|
||||
time_string(),
|
||||
epoch_str,
|
||||
valid_loss,
|
||||
valid_acc1,
|
||||
valid_acc5,
|
||||
valid_accuracies["best"],
|
||||
100 - valid_accuracies["best"],
|
||||
)
|
||||
)
|
||||
if valid_acc1 > valid_accuracies["best"]:
|
||||
valid_accuracies["best"] = valid_acc1
|
||||
find_best = True
|
||||
logger.log(
|
||||
"Currently, the best validation accuracy found at {:03d}-epoch :: acc@1={:.2f}, acc@5={:.2f}, error@1={:.2f}, error@5={:.2f}, save into {:}.".format(
|
||||
epoch,
|
||||
valid_acc1,
|
||||
valid_acc5,
|
||||
100 - valid_acc1,
|
||||
100 - valid_acc5,
|
||||
model_best_path,
|
||||
)
|
||||
)
|
||||
num_bytes = (
|
||||
torch.cuda.max_memory_cached(next(network.parameters()).device) * 1.0
|
||||
)
|
||||
logger.log(
|
||||
"[GPU-Memory-Usage on {:} is {:} bytes, {:.2f} KB, {:.2f} MB, {:.2f} GB.]".format(
|
||||
next(network.parameters()).device,
|
||||
int(num_bytes),
|
||||
num_bytes / 1e3,
|
||||
num_bytes / 1e6,
|
||||
num_bytes / 1e9,
|
||||
)
|
||||
)
|
||||
max_bytes[epoch] = num_bytes
|
||||
if epoch % 10 == 0:
|
||||
torch.cuda.empty_cache()
|
||||
optimizer.zero_grad()
|
||||
outputs = model(inputs)
|
||||
loss = objective(outputs, targets)
|
||||
|
||||
# save checkpoint
|
||||
save_path = save_checkpoint(
|
||||
{
|
||||
"epoch": epoch,
|
||||
"args": deepcopy(args),
|
||||
"max_bytes": deepcopy(max_bytes),
|
||||
"FLOP": flop,
|
||||
"PARAM": param,
|
||||
"valid_accuracies": deepcopy(valid_accuracies),
|
||||
"model-config": model_config._asdict(),
|
||||
"optim-config": optim_config._asdict(),
|
||||
"base-model": base_model.state_dict(),
|
||||
"scheduler": scheduler.state_dict(),
|
||||
"optimizer": optimizer.state_dict(),
|
||||
},
|
||||
model_base_path,
|
||||
logger,
|
||||
)
|
||||
if find_best:
|
||||
copy_checkpoint(model_base_path, model_best_path, logger)
|
||||
last_info = save_checkpoint(
|
||||
{
|
||||
"epoch": epoch,
|
||||
"args": deepcopy(args),
|
||||
"last_checkpoint": save_path,
|
||||
},
|
||||
logger.path("info"),
|
||||
logger,
|
||||
)
|
||||
loss.backward()
|
||||
optimizer.step()
|
||||
scheduler.step()
|
||||
if xiter % iters_per_epoch == 0:
|
||||
logger.log("TRAIN [{:}] loss = {:.6f}".format(iter_str, loss.item()))
|
||||
|
||||
# measure elapsed time
|
||||
epoch_time.update(time.time() - start_time)
|
||||
iter_time.update(time.time() - start_time)
|
||||
start_time = time.time()
|
||||
|
||||
logger.log("\n" + "-" * 200)
|
||||
logger.log(
|
||||
"Finish training/validation in {:} with Max-GPU-Memory of {:.2f} MB, and save final checkpoint into {:}".format(
|
||||
convert_secs2time(epoch_time.sum, True),
|
||||
max(v for k, v in max_bytes.items()) / 1e6,
|
||||
logger.path("info"),
|
||||
)
|
||||
)
|
||||
logger.log("-" * 200 + "\n")
|
||||
logger.close()
|
||||
|
||||
@@ -249,7 +135,7 @@ if __name__ == "__main__":
|
||||
parser.add_argument("--weight_decay", type=float, help="The weight decay")
|
||||
parser.add_argument("--scheduler", type=str, help="The scheduler indicator.")
|
||||
parser.add_argument("--steps", type=int, help="The total number of steps.")
|
||||
parser.add_argument("--batch_size", type=int, default=2, help="The batch size.")
|
||||
parser.add_argument("--batch_size", type=int, default=256, help="The batch size.")
|
||||
parser.add_argument("--workers", type=int, default=4, help="The number of workers")
|
||||
# Random Seed
|
||||
parser.add_argument("--rand_seed", type=int, default=-1, help="manual seed")
|
||||
|
||||
Reference in New Issue
Block a user