提交 3a9decba 编写于 作者: Z zhanghan17

fix multi process download model

上级 2ed67ee5
......@@ -21,6 +21,8 @@ import logging
from tqdm import tqdm
from pathlib import Path
import six
import paddle as P
import time
if six.PY2:
from pathlib2 import Path
else:
......@@ -33,6 +35,8 @@ def _fetch_from_remote(url,
force_download=False,
cached_dir='~/.paddle-ernie-cache'):
import hashlib, tempfile, requests, tarfile
env = P.distributed.ParallelEnv()
sig = hashlib.md5(url.encode('utf8')).hexdigest()
cached_dir = Path(cached_dir).expanduser()
try:
......@@ -40,7 +44,9 @@ def _fetch_from_remote(url,
except OSError:
pass
cached_dir_model = cached_dir / sig
if force_download or not cached_dir_model.exists():
done_file = cached_dir_model / 'fetch_done'
if force_download or not done_file.exists():
if env.dev_id == 0:
cached_dir_model.mkdir()
tmpfile = cached_dir_model / 'tmp'
with tmpfile.open('wb') as f:
......@@ -59,6 +65,15 @@ def _fetch_from_remote(url,
with tarfile.open(tmpfile.as_posix()) as tf:
tf.extractall(path=cached_dir_model.as_posix())
os.remove(tmpfile.as_posix())
f = done_file.open('wb')
f.close()
else:
while True:
if done_file.exists():
break
else:
time.sleep(1)
log.debug('%s cached in %s' % (url, cached_dir))
return cached_dir_model
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册