File size: 6,027 Bytes
789eef1 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 | # Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
import argparse
import os
import os.path as osp
from copy import deepcopy
from mmengine.config import Config, ConfigDict, DictAction
from mmengine.registry import RUNNERS
from mmengine.runner import Runner
from mmengine.utils import digit_version
from mmengine.utils.dl_utils import TORCH_VERSION
import torchvision
torchvision.disable_beta_transforms_warning()
def parse_args():
parser = argparse.ArgumentParser(description='Train a model')
parser.add_argument('config', help='train config file path')
parser.add_argument('--work-dir', help='the dir to save logs and models')
parser.add_argument(
'--resume',
nargs='?',
type=str,
const='auto',
help='If specify checkpoint path, resume from it, while if not '
'specify, try to auto resume from the latest checkpoint '
'in the work directory.')
parser.add_argument(
'--amp',
action='store_true',
help='enable automatic-mixed-precision training')
parser.add_argument(
'--no-validate',
action='store_true',
help='whether not to evaluate the checkpoint during training')
parser.add_argument(
'--auto-scale-lr',
action='store_true',
help='whether to auto scale the learning rate according to the '
'actual batch size and the original batch size.')
parser.add_argument(
'--no-pin-memory',
action='store_true',
help='whether to disable the pin_memory option in dataloaders.')
parser.add_argument(
'--no-persistent-workers',
action='store_true',
help='whether to disable the persistent_workers option in dataloaders.'
)
parser.add_argument(
'--cfg-options',
nargs='+',
action=DictAction,
help='override some settings in the used config, the key-value pair '
'in xxx=yyy format will be merged into config file. If the value to '
'be overwritten is a list, it should be like key="[a,b]" or key=a,b '
'It also allows nested list/tuple values, e.g. key="[(a,b),(c,d)]" '
'Note that the quotation marks are necessary and that no white space '
'is allowed.')
parser.add_argument(
'--launcher',
choices=['none', 'pytorch', 'slurm', 'mpi'],
default='none',
help='job launcher')
# When using PyTorch version >= 2.0.0, the `torch.distributed.launch`
# will pass the `--local-rank` parameter to `tools/train.py` instead
# of `--local_rank`.
parser.add_argument('--local_rank', '--local-rank', type=int, default=0)
args = parser.parse_args()
if 'LOCAL_RANK' not in os.environ:
os.environ['LOCAL_RANK'] = str(args.local_rank)
return args
def merge_args(cfg, args):
"""Merge CLI arguments to config."""
if args.no_validate:
cfg.val_cfg = None
cfg.val_dataloader = None
cfg.val_evaluator = None
cfg.launcher = args.launcher
# work_dir is determined in this priority: CLI > segment in file > filename
if args.work_dir is not None:
# update configs according to CLI args if args.work_dir is not None
cfg.work_dir = args.work_dir
elif cfg.get('work_dir', None) is None:
# use config filename as default work_dir if cfg.work_dir is None
cfg.work_dir = osp.join('./work_dirs',
osp.splitext(osp.basename(args.config))[0])
# enable automatic-mixed-precision training
if args.amp is True:
optim_wrapper = cfg.optim_wrapper.get('type', 'OptimWrapper')
assert optim_wrapper in ['OptimWrapper', 'AmpOptimWrapper'], \
'`--amp` is not supported custom optimizer wrapper type ' \
f'`{optim_wrapper}.'
cfg.optim_wrapper.type = 'AmpOptimWrapper'
cfg.optim_wrapper.setdefault('loss_scale', 'dynamic')
# resume training
if args.resume == 'auto':
cfg.resume = True
cfg.load_from = None
elif args.resume is not None:
cfg.resume = True
cfg.load_from = args.resume
# enable auto scale learning rate
if args.auto_scale_lr:
cfg.auto_scale_lr.enable = True
# set dataloader args
default_dataloader_cfg = ConfigDict(
pin_memory=True,
persistent_workers=True,
collate_fn=dict(type='default_collate'),
)
if digit_version(TORCH_VERSION) < digit_version('1.8.0'):
default_dataloader_cfg.persistent_workers = False
def set_default_dataloader_cfg(cfg, field):
if cfg.get(field, None) is None:
return
dataloader_cfg = deepcopy(default_dataloader_cfg)
dataloader_cfg.update(cfg[field])
cfg[field] = dataloader_cfg
if args.no_pin_memory:
cfg[field]['pin_memory'] = False
if args.no_persistent_workers:
cfg[field]['persistent_workers'] = False
set_default_dataloader_cfg(cfg, 'train_dataloader')
set_default_dataloader_cfg(cfg, 'val_dataloader')
set_default_dataloader_cfg(cfg, 'test_dataloader')
if args.cfg_options is not None:
cfg.merge_from_dict(args.cfg_options)
return cfg
def main():
args = parse_args()
# load config
cfg = Config.fromfile(args.config)
# merge cli arguments to config
cfg = merge_args(cfg, args)
# Set the train batch size to be read dataset class
os.environ['TRAIN_BATCH_SIZE_PER_GPU'] = str(cfg.train_dataloader.batch_size)
# build the runner from config
if 'runner_type' not in cfg:
# build the default runner
runner = Runner.from_cfg(cfg)
else:
# build customized runner from the registry
# if 'runner_type' is set in the cfg
runner = RUNNERS.build(cfg)
# start training
runner.train()
if __name__ == '__main__':
main()
|