mirrors/pytorch-image-models
1
0
Fork 0
mirror of https://github.com/huggingface/pytorch-image-models.git synced 2025-06-03 15:01:08 +08:00
Code Issues Projects Releases Wiki Activity

Make use of wandb configurable

Browse Source
This commit is contained in:
Aman Arora 2021-04-08 03:35:59 -04:00
parent 8e6fb861e4
commit 00c8e0b8bd
1 changed files with 17 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
train.py
View File

@ -273,6 +273,10 @@ parser.add_argument('--use-multi-epochs-loader', action='store_true', default=Fa
help='use the multi-epochs-loader to save time at the beginning of every epoch') help='use the multi-epochs-loader to save time at the beginning of every epoch')
parser.add_argument('--torchscript', dest='torchscript', action='store_true', parser.add_argument('--torchscript', dest='torchscript', action='store_true',
help='convert model torchscript for inference') help='convert model torchscript for inference')
parser.add_argument('--use-wandb', action='store_true', default=False,
help='use wandb for training and validation logs')
parser.add_argument('--wandb-project-name', type=str, default=None,
help='wandb project name to be used')
def _parse_args(): def _parse_args():
@ -295,8 +299,13 @@ def _parse_args():
def main(): def main():
setup_default_logging() setup_default_logging()
args, args_text = _parse_args() args, args_text = _parse_args()
wandb.init(project='efficientnet_v2', config=args)
wandb.run.name = args.model if args.use_wandb:
if not args.wandb_project_name:
args.wandb_project_name = args.model
_logger.warning(f"Wandb project name not provided, defaulting to {args.model}")
wandb.init(project=args.wandb_project_name, config=args)
args.prefetcher = not args.no_prefetcher args.prefetcher = not args.no_prefetcher
args.distributed = False args.distributed = False
if 'WORLD_SIZE' in os.environ: if 'WORLD_SIZE' in os.environ:
@ -575,14 +584,18 @@ def main():
epoch, model, loader_train, optimizer, train_loss_fn, args, epoch, model, loader_train, optimizer, train_loss_fn, args,
lr_scheduler=lr_scheduler, saver=saver, output_dir=output_dir, lr_scheduler=lr_scheduler, saver=saver, output_dir=output_dir,
amp_autocast=amp_autocast, loss_scaler=loss_scaler, model_ema=model_ema, mixup_fn=mixup_fn) amp_autocast=amp_autocast, loss_scaler=loss_scaler, model_ema=model_ema, mixup_fn=mixup_fn)
wandb.log(train_metrics)
if args.distributed and args.dist_bn in ('broadcast', 'reduce'): if args.distributed and args.dist_bn in ('broadcast', 'reduce'):
if args.local_rank == 0: if args.local_rank == 0:
_logger.info("Distributing BatchNorm running means and vars") _logger.info("Distributing BatchNorm running means and vars")
distribute_bn(model, args.world_size, args.dist_bn == 'reduce') distribute_bn(model, args.world_size, args.dist_bn == 'reduce')
eval_metrics = validate(model, loader_eval, validate_loss_fn, args, amp_autocast=amp_autocast) eval_metrics = validate(model, loader_eval, validate_loss_fn, args, amp_autocast=amp_autocast)
wandb.log(eval_metrics)
if args.use_wandb:
wandb.log(train_metrics)
wandb.log(eval_metrics)
if model_ema is not None and not args.model_ema_force_cpu: if model_ema is not None and not args.model_ema_force_cpu:
if args.distributed and args.dist_bn in ('broadcast', 'reduce'): if args.distributed and args.dist_bn in ('broadcast', 'reduce'):
distribute_bn(model_ema, args.world_size, args.dist_bn == 'reduce') distribute_bn(model_ema, args.world_size, args.dist_bn == 'reduce')