Skip to content

Support downloading multi data in modelart #275

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
May 11, 2023
Merged

Support downloading multi data in modelart #275

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e5d0f25
Support downloading multi data in modelart
zhtmike May 10, 2023
f0231fb
fix the default values for modelart
zhtmike May 10, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions tools/arg_parser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,13 +10,14 @@ def parse_args():
help='Run on modelarts platform (default=False)')
group.add_argument('--device_target', type=str, default='Ascend',
help='Target device, only used on modelarts platform (default=Ascend)')
group.add_argument('--multi_data_url', type=str, default='/cache/data/',
# The url are provided by modelart, usually they are S3 paths
group.add_argument('--multi_data_url', type=str, default='',
help='path to multi dataset')
group.add_argument('--data_url', type=str, default='/cache/data/',
group.add_argument('--data_url', type=str, default='',
help='path to dataset')
group.add_argument('--ckpt_url', type=str, default='/cache/output/',
group.add_argument('--ckpt_url', type=str, default='',
help='pre_train_model path in obs')
group.add_argument('--train_url', type=str, default='/cache/output/',
group.add_argument('--train_url', type=str, default='',
help='model folder to save/load')

args = parser.parse_args()
Expand Down
9 changes: 8 additions & 1 deletion tools/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -207,11 +207,18 @@ def main(cfg):

if args.enable_modelarts:
import moxing as mox
from ast import literal_eval
from tools.modelarts_adapter.modelarts import get_device_id, sync_data

dataset_root = '/cache/data/'
# download dataset from server to local on device 0, other devices will wait until data sync finished.
sync_data(args.data_url, dataset_root)
if args.multi_data_url:
multi_data_url = literal_eval(args.multi_data_url)
for x in multi_data_url:
sync_data(x['dataset_url'], dataset_root)
else:
sync_data(args.data_url, dataset_root)

if get_device_id() == 0:
# mox.file.copy_parallel(src_url=args.data_url, dst_url=dataset_root)
print(f'INFO: datasets found: {os.listdir(dataset_root)} \n'
Expand Down