Unanswered
Hi Everyone! I Would Need Your Opinion About How To Proceed With Offline Mode When Serving Models. We Currently Have Some Serving Endpoints Which Are Constantly Running And We Report Images And Statistics. How Do You Recommend Dealing With Offline Mode In
Hi WickedWhale51 ! ClearML is tolerant to network failures. Anyway, if you wish the upload the offline data periodically, you could zip the offline mode folder and import it:
# make sure the state of the offline data is saved
Task.current_task()._edit()
# create zip file
offline_folder = Task.current_task().get_offline_mode_folder()
zip_file = offline_folder.as_posix() + ".zip"
with ZipFile(zip_file, "w", allowZip64=True, compression=ZIP_DEFLATED) as zf:
for filename in offline_folder.rglob("*"):
if filename.is_file():
relative_file_name = filename.relative_to(offline_folder).as_posix()
zf.write(filename.as_posix(), arcname=relative_file_name)
Task.import_offline_session(zip_file, task_holding_reports="PASS_VALID_ID_TO_AVOID_DUPLICATES")
Also, note the task_holding_reports
argument
7 Views
0
Answers
10 days ago
10 days ago