[Feature] Add diff seeds to diff ranks and set torch seed in worker_i…

…nit_fn (open-mmlab#113)

* add init_random_seed

* Set diff seed to diff workers

mmrazor/apis/__init__.py

@@ -2,4 +2,6 @@
 from .mmcls import *  # noqa: F401,F403
 from .mmdet import *  # noqa: F401,F403
 from .mmseg import *  # noqa: F401,F403
-from .utils import set_random_seed  # noqa: F401
+from .utils import init_random_seed, set_random_seed  # noqa: F401
+
+__all__ = ['init_random_seed', 'set_random_seed']

mmrazor/apis/utils.py

@@ -3,6 +3,39 @@
 
 import numpy as np
 import torch
+import torch.distributed as dist
+from mmcv.runner import get_dist_info
+
+
+def init_random_seed(seed=None, device='cuda'):
+    """Initialize random seed.
+
+    If the seed is not set, the seed will be automatically randomized,
+    and then broadcast to all processes to prevent some potential bugs.
+    Args:
+        seed (int, Optional): The seed. Default to None.
+        device (str): The device where the seed will be put on.
+            Default to 'cuda'.
+    Returns:
+        int: Seed to be used.
+    """
+    if seed is not None:
+        return seed
+
+    # Make sure all ranks share the same random seed to prevent
+    # some potential bugs. Please refer to
+    # https://github.com/open-mmlab/mmdetection/issues/6339
+    rank, world_size = get_dist_info()
+    seed = np.random.randint(2**31)
+    if world_size == 1:
+        return seed
+
+    if rank == 0:
+        random_num = torch.tensor(seed, dtype=torch.int32, device=device)
+    else:
+        random_num = torch.tensor(0, dtype=torch.int32, device=device)
+    dist.broadcast(random_num, src=0)
+    return random_num.item()
 
 
 def set_random_seed(seed: int, deterministic: bool = False) -> None:

tools/mmcls/train_mmcls.py

@@ -8,14 +8,15 @@
 
 import mmcv
 import torch
+import torch.distributed as dist
 from mmcls import __version__
 from mmcls.datasets import build_dataset
 from mmcls.utils import collect_env, get_root_logger
 from mmcv import Config, DictAction
 from mmcv.runner import get_dist_info, init_dist
 
 # Differences from mmclassification
-from mmrazor.apis import set_random_seed, train_mmcls_model
+from mmrazor.apis import init_random_seed, set_random_seed, train_mmcls_model
 from mmrazor.models import build_algorithm
 from mmrazor.utils import setup_multi_processes
 
@@ -54,6 +55,10 @@ def parse_args():
         help='id of gpu to use '
         '(only applicable to non-distributed training)')
     parser.add_argument('--seed', type=int, default=None, help='random seed')
+    parser.add_argument(
+        '--diff_seed',
+        action='store_true',
+        help='Whether or not set different seeds for different ranks')
     parser.add_argument(
         '--deterministic',
         action='store_true',
@@ -154,12 +159,14 @@ def main():
     logger.info(f'Config:\n{cfg.pretty_text}')
 
     # set random seeds
-    if args.seed is not None:
-        logger.info(f'Set random seed to {args.seed}, '
-                    f'deterministic: {args.deterministic}')
-        set_random_seed(args.seed, deterministic=args.deterministic)
-    cfg.seed = args.seed
-    meta['seed'] = args.seed
+    seed = init_random_seed(args.seed)
+    seed = seed + dist.get_rank() if args.diff_seed else seed
+    logger.info(f'Set random seed to {seed}, '
+                f'deterministic: {args.deterministic}')
+    set_random_seed(seed, deterministic=args.deterministic)
+    cfg.seed = seed
+    meta['seed'] = seed
+    meta['exp_name'] = osp.basename(args.config)
 
     # Difference from mmclassification
     # replace `model` to `algorithm`

tools/mmdet/train_mmdet.py

@@ -16,14 +16,15 @@
 
 import mmcv
 import torch
+import torch.distributed as dist
 from mmcv import Config, DictAction
 from mmcv.runner import get_dist_info, init_dist
 from mmcv.utils import get_git_hash
 from mmdet import __version__
 from mmdet.datasets import build_dataset
 from mmdet.utils import collect_env, get_root_logger
 
-from mmrazor.apis import set_random_seed, train_mmdet_model
+from mmrazor.apis import init_random_seed, set_random_seed, train_mmdet_model
 from mmrazor.models import build_algorithm
 from mmrazor.utils import setup_multi_processes
 
@@ -61,6 +62,10 @@ def parse_args():
         help='id of gpu to use '
         '(only applicable to non-distributed training)')
     parser.add_argument('--seed', type=int, default=None, help='random seed')
+    parser.add_argument(
+        '--diff_seed',
+        action='store_true',
+        help='Whether or not set different seeds for different ranks')
     parser.add_argument(
         '--deterministic',
         action='store_true',
@@ -166,12 +171,13 @@ def main():
     logger.info(f'Config:\n{cfg.pretty_text}')
 
     # set random seeds
-    if args.seed is not None:
-        logger.info(f'Set random seed to {args.seed}, '
-                    f'deterministic: {args.deterministic}')
-        set_random_seed(args.seed, deterministic=args.deterministic)
-    cfg.seed = args.seed
-    meta['seed'] = args.seed
+    seed = init_random_seed(args.seed)
+    seed = seed + dist.get_rank() if args.diff_seed else seed
+    logger.info(f'Set random seed to {seed}, '
+                f'deterministic: {args.deterministic}')
+    set_random_seed(seed, deterministic=args.deterministic)
+    cfg.seed = seed
+    meta['seed'] = seed
     meta['exp_name'] = osp.basename(args.config)
 
     algorithm = build_algorithm(cfg.algorithm)

tools/mmseg/train_mmseg.py

@@ -16,6 +16,7 @@
 
 import mmcv
 import torch
+import torch.distributed as dist
 from mmcv.cnn.utils import revert_sync_batchnorm
 from mmcv.runner import get_dist_info, init_dist
 from mmcv.utils import Config, DictAction, get_git_hash
@@ -24,7 +25,7 @@
 from mmseg.utils import collect_env, get_root_logger
 
 # Differences from mmdetection
-from mmrazor.apis import set_random_seed, train_mmseg_model
+from mmrazor.apis import init_random_seed, set_random_seed, train_mmseg_model
 from mmrazor.models.builder import build_algorithm
 from mmrazor.utils import setup_multi_processes
 
@@ -64,6 +65,10 @@ def parse_args():
         help='id of gpu to use '
         '(only applicable to non-distributed training)')
     parser.add_argument('--seed', type=int, default=None, help='random seed')
+    parser.add_argument(
+        '--diff_seed',
+        action='store_true',
+        help='Whether or not set different seeds for different ranks')
     parser.add_argument(
         '--deterministic',
         action='store_true',
@@ -168,11 +173,12 @@ def main():
     logger.info(f'Config:\n{cfg.pretty_text}')
 
     # set random seeds
-    if args.seed is not None:
-        logger.info(f'Set random seed to {args.seed}, deterministic: '
-                    f'{args.deterministic}')
-        set_random_seed(args.seed, deterministic=args.deterministic)
-    cfg.seed = args.seed
+    seed = init_random_seed(args.seed)
+    seed = seed + dist.get_rank() if args.diff_seed else seed
+    logger.info(f'Set random seed to {args.seed}, deterministic: '
+                f'{args.deterministic}')
+    set_random_seed(args.seed, deterministic=args.deterministic)
+    cfg.seed = seed
     meta['seed'] = args.seed
     meta['exp_name'] = osp.basename(args.config)