[fbsync] Warmup schedulers in References (#4411)

Summary:
* Warmup on Classficiation references.

* Adjust epochs for cosine.

* Warmup on Segmentation references.

* Warmup on Video classification references.

* Adding support of both types of warmup in segmentation.

* Use LinearLR in detection.

* Fix deprecation warning.

Reviewed By: datumbox

Differential Revision: D31268039

fbshipit-source-id: d0fe7e334c01201c2413bac8b911d740b9a69bba

Author: NicolasHug

references/classification/train.py

@@ -208,7 +208,25 @@ def main(args):
                                           opt_level=args.apex_opt_level
                                           )
 
-    lr_scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=args.lr_step_size, gamma=args.lr_gamma)
+    args.lr_scheduler = args.lr_scheduler.lower()
+    if args.lr_scheduler == 'steplr':
+        main_lr_scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=args.lr_step_size, gamma=args.lr_gamma)
+    elif args.lr_scheduler == 'cosineannealinglr':
+        main_lr_scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer,
+                                                                       T_max=args.epochs - args.lr_warmup_epochs)
+    else:
+        raise RuntimeError("Invalid lr scheduler '{}'. Only StepLR and CosineAnnealingLR "
+                           "are supported.".format(args.lr_scheduler))
+
+    if args.lr_warmup_epochs > 0:
+        lr_scheduler = torch.optim.lr_scheduler.SequentialLR(
+            optimizer,
+            schedulers=[torch.optim.lr_scheduler.ConstantLR(optimizer, factor=args.lr_warmup_decay,
+                                                            total_iters=args.lr_warmup_epochs), main_lr_scheduler],
+            milestones=[args.lr_warmup_epochs]
+        )
+    else:
+        lr_scheduler = main_lr_scheduler
 
     model_without_ddp = model
     if args.distributed:
@@ -287,6 +305,9 @@ def get_args_parser(add_help=True):
                         dest='label_smoothing')
     parser.add_argument('--mixup-alpha', default=0.0, type=float, help='mixup alpha (default: 0.0)')
     parser.add_argument('--cutmix-alpha', default=0.0, type=float, help='cutmix alpha (default: 0.0)')
+    parser.add_argument('--lr-scheduler', default="steplr", help='the lr scheduler (default: steplr)')
+    parser.add_argument('--lr-warmup-epochs', default=0, type=int, help='the number of epochs to warmup (default: 0)')
+    parser.add_argument('--lr-warmup-decay', default=0.01, type=int, help='the decay for lr')
     parser.add_argument('--lr-step-size', default=30, type=int, help='decrease lr every step-size epochs')
     parser.add_argument('--lr-gamma', default=0.1, type=float, help='decrease lr by a factor of lr-gamma')
     parser.add_argument('--print-freq', default=10, type=int, help='print frequency')

references/detection/engine.py

@@ -21,7 +21,8 @@ def train_one_epoch(model, optimizer, data_loader, device, epoch, print_freq):
         warmup_factor = 1. / 1000
         warmup_iters = min(1000, len(data_loader) - 1)
 
-        lr_scheduler = utils.warmup_lr_scheduler(optimizer, warmup_iters, warmup_factor)
+        lr_scheduler = torch.optim.lr_scheduler.LinearLR(optimizer, start_factor=warmup_factor,
+                                                         total_iters=warmup_iters)
 
     for images, targets in metric_logger.log_every(data_loader, print_freq, header):
         images = list(image.to(device) for image in images)

references/detection/utils.py

@@ -207,17 +207,6 @@ def collate_fn(batch):
     return tuple(zip(*batch))
 
 
-def warmup_lr_scheduler(optimizer, warmup_iters, warmup_factor):
-
-    def f(x):
-        if x >= warmup_iters:
-            return 1
-        alpha = float(x) / warmup_iters
-        return warmup_factor * (1 - alpha) + alpha
-
-    return torch.optim.lr_scheduler.LambdaLR(optimizer, f)
-
-
 def mkdir(path):
     try:
         os.makedirs(path)

references/segmentation/train.py

@@ -133,9 +133,30 @@ def main(args):
         params_to_optimize,
         lr=args.lr, momentum=args.momentum, weight_decay=args.weight_decay)
 
-    lr_scheduler = torch.optim.lr_scheduler.LambdaLR(
+    iters_per_epoch = len(data_loader)
+    main_lr_scheduler = torch.optim.lr_scheduler.LambdaLR(
         optimizer,
-        lambda x: (1 - x / (len(data_loader) * args.epochs)) ** 0.9)
+        lambda x: (1 - x / (iters_per_epoch * (args.epochs - args.lr_warmup_epochs))) ** 0.9)
+
+    if args.lr_warmup_epochs > 0:
+        warmup_iters = iters_per_epoch * args.lr_warmup_epochs
+        args.lr_warmup_method = args.lr_warmup_method.lower()
+        if args.lr_warmup_method == 'linear':
+            warmup_lr_scheduler = torch.optim.lr_scheduler.LinearLR(optimizer, start_factor=args.lr_warmup_decay,
+                                                                    total_iters=warmup_iters)
+        elif args.lr_warmup_method == 'constant':
+            warmup_lr_scheduler = torch.optim.lr_scheduler.ConstantLR(optimizer, factor=args.lr_warmup_decay,
+                                                                      total_iters=warmup_iters)
+        else:
+            raise RuntimeError("Invalid warmup lr method '{}'. Only linear and constant "
+                               "are supported.".format(args.lr_warmup_method))
+        lr_scheduler = torch.optim.lr_scheduler.SequentialLR(
+            optimizer,
+            schedulers=[warmup_lr_scheduler, main_lr_scheduler],
+            milestones=[warmup_iters]
+        )
+    else:
+        lr_scheduler = main_lr_scheduler
 
     if args.resume:
         checkpoint = torch.load(args.resume, map_location='cpu')
@@ -197,6 +218,9 @@ def get_args_parser(add_help=True):
     parser.add_argument('--wd', '--weight-decay', default=1e-4, type=float,
                         metavar='W', help='weight decay (default: 1e-4)',
                         dest='weight_decay')
+    parser.add_argument('--lr-warmup-epochs', default=0, type=int, help='the number of epochs to warmup (default: 0)')
+    parser.add_argument('--lr-warmup-method', default="linear", type=str, help='the warmup method (default: linear)')
+    parser.add_argument('--lr-warmup-decay', default=0.01, type=int, help='the decay for lr')
     parser.add_argument('--print-freq', default=10, type=int, help='print frequency')
     parser.add_argument('--output-dir', default='.', help='path where to save')
     parser.add_argument('--resume', default='', help='resume from checkpoint')

references/segmentation/transforms.py

@@ -37,7 +37,7 @@ def __init__(self, min_size, max_size=None):
     def __call__(self, image, target):
         size = random.randint(self.min_size, self.max_size)
         image = F.resize(image, size)
-        target = F.resize(target, size, interpolation=Image.NEAREST)
+        target = F.resize(target, size, interpolation=T.InterpolationMode.NEAREST)
         return image, target
 
 

references/video_classification/scheduler.py

@@ -12,8 +12,6 @@
 import presets
 import utils
 
-from scheduler import WarmupMultiStepLR
-
 try:
     from apex import amp
 except ImportError:
@@ -202,11 +200,30 @@ def main(args):
 
     # convert scheduler to be per iteration, not per epoch, for warmup that lasts
     # between different epochs
-    warmup_iters = args.lr_warmup_epochs * len(data_loader)
-    lr_milestones = [len(data_loader) * m for m in args.lr_milestones]
-    lr_scheduler = WarmupMultiStepLR(
-        optimizer, milestones=lr_milestones, gamma=args.lr_gamma,
-        warmup_iters=warmup_iters, warmup_factor=1e-5)
+    iters_per_epoch = len(data_loader)
+    lr_milestones = [iters_per_epoch * (m - args.lr_warmup_epochs) for m in args.lr_milestones]
+    main_lr_scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, milestones=lr_milestones, gamma=args.lr_gamma)
+
+    if args.lr_warmup_epochs > 0:
+        warmup_iters = iters_per_epoch * args.lr_warmup_epochs
+        args.lr_warmup_method = args.lr_warmup_method.lower()
+        if args.lr_warmup_method == 'linear':
+            warmup_lr_scheduler = torch.optim.lr_scheduler.LinearLR(optimizer, start_factor=args.lr_warmup_decay,
+                                                                    total_iters=warmup_iters)
+        elif args.lr_warmup_method == 'constant':
+            warmup_lr_scheduler = torch.optim.lr_scheduler.ConstantLR(optimizer, factor=args.lr_warmup_decay,
+                                                                      total_iters=warmup_iters)
+        else:
+            raise RuntimeError("Invalid warmup lr method '{}'. Only linear and constant "
+                               "are supported.".format(args.lr_warmup_method))
+
+        lr_scheduler = torch.optim.lr_scheduler.SequentialLR(
+            optimizer,
+            schedulers=[warmup_lr_scheduler, main_lr_scheduler],
+            milestones=[warmup_iters]
+        )
+    else:
+        lr_scheduler = main_lr_scheduler
 
     model_without_ddp = model
     if args.distributed:
@@ -277,7 +294,9 @@ def parse_args():
                         dest='weight_decay')
     parser.add_argument('--lr-milestones', nargs='+', default=[20, 30, 40], type=int, help='decrease lr on milestones')
     parser.add_argument('--lr-gamma', default=0.1, type=float, help='decrease lr by a factor of lr-gamma')
-    parser.add_argument('--lr-warmup-epochs', default=10, type=int, help='number of warmup epochs')
+    parser.add_argument('--lr-warmup-epochs', default=10, type=int, help='the number of epochs to warmup (default: 10)')
+    parser.add_argument('--lr-warmup-method', default="linear", type=str, help='the warmup method (default: linear)')
+    parser.add_argument('--lr-warmup-decay', default=0.001, type=int, help='the decay for lr')
     parser.add_argument('--print-freq', default=10, type=int, help='print frequency')
     parser.add_argument('--output-dir', default='.', help='path where to save')
     parser.add_argument('--resume', default='', help='resume from checkpoint')