Reduce variance of evaluation in reference

references/detection/train.py

@@ -132,6 +132,10 @@ def get_args_parser(add_help=True):
         action="store_true",
     )
 
+    parser.add_argument(
+        "--use-deterministic-algorithms", action="store_true", help="Forces the use of deterministic algorithms only."
+    )
+
     # distributed training parameters
     parser.add_argument("--world-size", default=1, type=int, help="number of distributed processes")
     parser.add_argument("--dist-url", default="env://", type=str, help="url used to set up distributed training")
@@ -153,6 +157,12 @@ def main(args):
 
     device = torch.device(args.device)
 
+    if args.use_deterministic_algorithms:
+        torch.backends.cudnn.benchmark = False
+        torch.use_deterministic_algorithms(True)
+    else:
+        torch.backends.cudnn.benchmark = True
+
     # Data loading code
     print("Loading data")
 
@@ -162,7 +172,7 @@ def main(args):
     print("Creating data loaders")
     if args.distributed:
         train_sampler = torch.utils.data.distributed.DistributedSampler(dataset)
-        test_sampler = torch.utils.data.distributed.DistributedSampler(dataset_test)
+        test_sampler = torch.utils.data.distributed.DistributedSampler(dataset_test, shuffle=False)
     else:
         train_sampler = torch.utils.data.RandomSampler(dataset)
         test_sampler = torch.utils.data.SequentialSampler(dataset_test)
@@ -243,6 +253,9 @@ def main(args):
             scaler.load_state_dict(checkpoint["scaler"])
 
     if args.test_only:
+        # We disable the cudnn benchmarking because it can noticeably affect the accuracy
+        torch.backends.cudnn.benchmark = False
+        torch.backends.cudnn.deterministic = True
         evaluate(model, data_loader_test, device=device)
         return
 

references/optical_flow/train.py

@@ -209,6 +209,12 @@ def main(args):
         raise ValueError("The device must be cuda if we want to run in distributed mode using torchrun")
     device = torch.device(args.device)
 
+    if args.use_deterministic_algorithms:
+        torch.backends.cudnn.benchmark = False
+        torch.use_deterministic_algorithms(True)
+    else:
+        torch.backends.cudnn.benchmark = True
+
     model = torchvision.models.optical_flow.__dict__[args.model](weights=args.weights)
 
     if args.distributed:
@@ -370,6 +376,9 @@ def get_args_parser(add_help=True):
 
     parser.add_argument("--weights", default=None, type=str, help="the weights enum name to load.")
     parser.add_argument("--device", default="cuda", type=str, help="device (Use cuda or cpu, Default: cuda)")
+    parser.add_argument(
+        "--use-deterministic-algorithms", action="store_true", help="Forces the use of deterministic algorithms only."
+    )
 
     return parser
 

references/segmentation/train.py

@@ -1,6 +1,7 @@
 import datetime
 import os
 import time
+import warnings
 
 import presets
 import torch
@@ -61,16 +62,34 @@ def evaluate(model, data_loader, device, num_classes):
     confmat = utils.ConfusionMatrix(num_classes)
     metric_logger = utils.MetricLogger(delimiter="  ")
     header = "Test:"
+    num_processed_samples = 0
     with torch.inference_mode():
         for image, target in metric_logger.log_every(data_loader, 100, header):
             image, target = image.to(device), target.to(device)
             output = model(image)
             output = output["out"]
 
             confmat.update(target.flatten(), output.argmax(1).flatten())
+            # FIXME need to take into account that the datasets
+            # could have been padded in distributed setup
+            num_processed_samples += image.shape[0]
 
         confmat.reduce_from_all_processes()
 
+    num_processed_samples = utils.reduce_across_processes(num_processed_samples)
+    if (
+        hasattr(data_loader.dataset, "__len__")
+        and len(data_loader.dataset) != num_processed_samples
+        and torch.distributed.get_rank() == 0
+    ):
+        # See FIXME above
+        warnings.warn(
+            f"It looks like the dataset has {len(data_loader.dataset)} samples, but {num_processed_samples} "
+            "samples were used for the validation, which might bias the results. "
+            "Try adjusting the batch size and / or the world size. "
+            "Setting the world size to 1 is always a safe bet."
+        )
+
     return confmat
 
 
@@ -108,12 +127,18 @@ def main(args):
 
     device = torch.device(args.device)
 
+    if args.use_deterministic_algorithms:
+        torch.backends.cudnn.benchmark = False
+        torch.use_deterministic_algorithms(True)
+    else:
+        torch.backends.cudnn.benchmark = True
+
     dataset, num_classes = get_dataset(args.data_path, args.dataset, "train", get_transform(True, args))
     dataset_test, _ = get_dataset(args.data_path, args.dataset, "val", get_transform(False, args))
 
     if args.distributed:
         train_sampler = torch.utils.data.distributed.DistributedSampler(dataset)
-        test_sampler = torch.utils.data.distributed.DistributedSampler(dataset_test)
+        test_sampler = torch.utils.data.distributed.DistributedSampler(dataset_test, shuffle=False)
     else:
         train_sampler = torch.utils.data.RandomSampler(dataset)
         test_sampler = torch.utils.data.SequentialSampler(dataset_test)
@@ -191,6 +216,9 @@ def main(args):
                 scaler.load_state_dict(checkpoint["scaler"])
 
     if args.test_only:
+        # We disable the cudnn benchmarking because it can noticeably affect the accuracy
+        torch.backends.cudnn.benchmark = False
+        torch.backends.cudnn.deterministic = True
         confmat = evaluate(model, data_loader_test, device=device, num_classes=num_classes)
         print(confmat)
         return
@@ -261,6 +289,9 @@ def get_args_parser(add_help=True):
         help="Only test the model",
         action="store_true",
     )
+    parser.add_argument(
+        "--use-deterministic-algorithms", action="store_true", help="Forces the use of deterministic algorithms only."
+    )
     # distributed training parameters
     parser.add_argument("--world-size", default=1, type=int, help="number of distributed processes")
     parser.add_argument("--dist-url", default="env://", type=str, help="url used to set up distributed training")

references/segmentation/utils.py

@@ -30,11 +30,7 @@ def synchronize_between_processes(self):
         """
         Warning: does not synchronize the deque!
         """
-        if not is_dist_avail_and_initialized():
-            return
-        t = torch.tensor([self.count, self.total], dtype=torch.float64, device="cuda")
-        dist.barrier()
-        dist.all_reduce(t)
+        t = reduce_across_processes([self.count, self.total])
         t = t.tolist()
         self.count = int(t[0])
         self.total = t[1]
@@ -92,12 +88,7 @@ def compute(self):
         return acc_global, acc, iu
 
     def reduce_from_all_processes(self):
-        if not torch.distributed.is_available():
-            return
-        if not torch.distributed.is_initialized():
-            return
-        torch.distributed.barrier()
-        torch.distributed.all_reduce(self.mat)
+        reduce_across_processes(self.mat)
 
     def __str__(self):
         acc_global, acc, iu = self.compute()
@@ -296,3 +287,14 @@ def init_distributed_mode(args):
     )
     torch.distributed.barrier()
     setup_for_distributed(args.rank == 0)
+
+
+def reduce_across_processes(val):
+    if not is_dist_avail_and_initialized():
+        # nothing to sync, but we still convert to tensor for consistency with the distributed case.
+        return torch.tensor(val)
+
+    t = torch.tensor(val, device="cuda")
+    dist.barrier()
+    dist.all_reduce(t)
+    return t

references/similarity/train.py

@@ -88,6 +88,13 @@ def save(model, epoch, save_dir, file_name):
 
 def main(args):
     device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+
+    if args.use_deterministic_algorithms:
+        torch.backends.cudnn.benchmark = False
+        torch.use_deterministic_algorithms(True)
+    else:
+        torch.backends.cudnn.benchmark = True
+
     p = args.labels_per_batch
     k = args.samples_per_label
     batch_size = p * k
@@ -126,6 +133,13 @@ def main(args):
     )
     test_loader = DataLoader(test_dataset, batch_size=args.eval_batch_size, shuffle=False, num_workers=args.workers)
 
+    if args.test_only:
+        # We disable the cudnn benchmarking because it can noticeably affect the accuracy
+        torch.backends.cudnn.benchmark = False
+        torch.backends.cudnn.deterministic = True
+        evaluate(model, test_loader, device)
+        return
+
     for epoch in range(1, args.epochs + 1):
         print("Training...")
         train_epoch(model, optimizer, criterion, train_loader, device, epoch, args.print_freq)
@@ -155,6 +169,15 @@ def parse_args():
     parser.add_argument("--print-freq", default=20, type=int, help="print frequency")
     parser.add_argument("--save-dir", default=".", type=str, help="Model save directory")
     parser.add_argument("--resume", default="", type=str, help="path of checkpoint")
+    parser.add_argument(
+        "--test-only",
+        dest="test_only",
+        help="Only test the model",
+        action="store_true",
+    )
+    parser.add_argument(
+        "--use-deterministic-algorithms", action="store_true", help="Forces the use of deterministic algorithms only."
+    )
 
     return parser.parse_args()
 

references/video_classification/train.py

@@ -1,6 +1,7 @@
 import datetime
 import os
 import time
+import warnings
 
 import presets
 import torch
@@ -50,6 +51,7 @@ def evaluate(model, criterion, data_loader, device):
     model.eval()
     metric_logger = utils.MetricLogger(delimiter="  ")
     header = "Test:"
+    num_processed_samples = 0
     with torch.inference_mode():
         for video, target in metric_logger.log_every(data_loader, 100, header):
             video = video.to(device, non_blocking=True)
@@ -64,7 +66,28 @@ def evaluate(model, criterion, data_loader, device):
             metric_logger.update(loss=loss.item())
             metric_logger.meters["acc1"].update(acc1.item(), n=batch_size)
             metric_logger.meters["acc5"].update(acc5.item(), n=batch_size)
+            num_processed_samples += batch_size
     # gather the stats from all processes
+    num_processed_samples = utils.reduce_across_processes(num_processed_samples)
+    if isinstance(data_loader.sampler, DistributedSampler):
+        # Get the len of UniformClipSampler inside DistributedSampler
+        num_data_from_sampler = len(data_loader.sampler.dataset)
+    else:
+        num_data_from_sampler = len(data_loader.sampler)
+
+    if (
+        hasattr(data_loader.dataset, "__len__")
+        and num_data_from_sampler != num_processed_samples
+        and torch.distributed.get_rank() == 0
+    ):
+        # See FIXME above
+        warnings.warn(
+            f"It looks like the sampler has {num_data_from_sampler} samples, but {num_processed_samples} "
+            "samples were used for the validation, which might bias the results. "
+            "Try adjusting the batch size and / or the world size. "
+            "Setting the world size to 1 is always a safe bet."
+        )
+
     metric_logger.synchronize_between_processes()
 
     print(
@@ -99,7 +122,11 @@ def main(args):
 
     device = torch.device(args.device)
 
-    torch.backends.cudnn.benchmark = True
+    if args.use_deterministic_algorithms:
+        torch.backends.cudnn.benchmark = False
+        torch.use_deterministic_algorithms(True)
+    else:
+        torch.backends.cudnn.benchmark = True
 
     # Data loading code
     print("Loading data")
@@ -173,7 +200,7 @@ def main(args):
     test_sampler = UniformClipSampler(dataset_test.video_clips, args.clips_per_video)
     if args.distributed:
         train_sampler = DistributedSampler(train_sampler)
-        test_sampler = DistributedSampler(test_sampler)
+        test_sampler = DistributedSampler(test_sampler, shuffle=False)
 
     data_loader = torch.utils.data.DataLoader(
         dataset,
@@ -248,6 +275,9 @@ def main(args):
             scaler.load_state_dict(checkpoint["scaler"])
 
     if args.test_only:
+        # We disable the cudnn benchmarking because it can noticeably affect the accuracy
+        torch.backends.cudnn.benchmark = False
+        torch.backends.cudnn.deterministic = True
         evaluate(model, criterion, data_loader_test, device=device)
         return
 
@@ -335,6 +365,9 @@ def parse_args():
         help="Only test the model",
         action="store_true",
     )
+    parser.add_argument(
+        "--use-deterministic-algorithms", action="store_true", help="Forces the use of deterministic algorithms only."
+    )
 
     # distributed training parameters
     parser.add_argument("--world-size", default=1, type=int, help="number of distributed processes")

references/video_classification/utils.py

@@ -30,11 +30,7 @@ def synchronize_between_processes(self):
         """
         Warning: does not synchronize the deque!
         """
-        if not is_dist_avail_and_initialized():
-            return
-        t = torch.tensor([self.count, self.total], dtype=torch.float64, device="cuda")
-        dist.barrier()
-        dist.all_reduce(t)
+        t = reduce_across_processes([self.count, self.total])
         t = t.tolist()
         self.count = int(t[0])
         self.total = t[1]
@@ -255,3 +251,14 @@ def init_distributed_mode(args):
     )
     torch.distributed.barrier()
     setup_for_distributed(args.rank == 0)
+
+
+def reduce_across_processes(val):
+    if not is_dist_avail_and_initialized():
+        # nothing to sync, but we still convert to tensor for consistency with the distributed case.
+        return torch.tensor(val)
+
+    t = torch.tensor(val, device="cuda")
+    dist.barrier()
+    dist.all_reduce(t)
+    return t