ClearML/clearml
1
0
Fork 0
mirror of https://github.com/clearml/clearml synced 2025-02-07 13:23:40 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix pipeline should always use artifacts hash for checking cache hits

Browse Source
This commit is contained in:
allegroai 2023-08-04 19:06:15 +03:00
parent 46c6d2bf0f
commit 99c7eecbee
1 changed files with 18 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
clearml/automation/job.py
View File

@ -384,10 +384,9 @@ class BaseJob(object):
section_overrides=None, section_overrides=None,
params_override=None, params_override=None,
configurations_override=None, configurations_override=None,
explicit_docker_image=None, explicit_docker_image=None
account_for_artifacts_hashes=True
): ):
# type: (Task, Optional[dict], Optional[dict], Optional[dict], Optional[str], bool) -> Optional[str] # type: (Task, Optional[dict], Optional[dict], Optional[dict], Optional[str]) -> Optional[str]
""" """
Create Hash (str) representing the state of the Task Create Hash (str) representing the state of the Task
@ -398,8 +397,6 @@ class BaseJob(object):
:param configurations_override: dictionary of configuration override objects (tasks.ConfigurationItem) :param configurations_override: dictionary of configuration override objects (tasks.ConfigurationItem)
:param explicit_docker_image: The explicit docker image. Used to invalidate the hash when the docker image :param explicit_docker_image: The explicit docker image. Used to invalidate the hash when the docker image
was explicitly changed was explicitly changed
:param account_for_artifacts_hashes: Calculate the hash of the task by accounting for the hashes of the
artifacts in `kwargs_artifacts` (as opposed of the task ID/artifact name stored in this section)
:return: str hash of the Task configuration :return: str hash of the Task configuration
""" """
@ -420,7 +417,6 @@ class BaseJob(object):
script.pop("requirements", None) script.pop("requirements", None)
hyper_params = deepcopy(task.get_parameters() if params_override is None else params_override) hyper_params = deepcopy(task.get_parameters() if params_override is None else params_override)
if account_for_artifacts_hashes:
hyper_params_to_change = {} hyper_params_to_change = {}
task_cache = {} task_cache = {}
for key, value in hyper_params.items(): for key, value in hyper_params.items():
@ -604,14 +600,6 @@ class ClearmlJob(BaseJob):
if allow_caching: if allow_caching:
# look for a cached copy of the Task # look for a cached copy of the Task
# get parameters + task_overrides + as dict and hash it. # get parameters + task_overrides + as dict and hash it.
task_hash_legacy = self._create_task_hash(
base_temp_task,
section_overrides=sections,
params_override=task_params,
configurations_override=configuration_overrides or None,
explicit_docker_image=kwargs.get("explicit_docker_image"),
account_for_artifacts_hashes=False
)
task_hash = self._create_task_hash( task_hash = self._create_task_hash(
base_temp_task, base_temp_task,
section_overrides=sections, section_overrides=sections,
@ -619,7 +607,7 @@ class ClearmlJob(BaseJob):
configurations_override=configuration_overrides or None, configurations_override=configuration_overrides or None,
explicit_docker_image=kwargs.get("explicit_docker_image") explicit_docker_image=kwargs.get("explicit_docker_image")
) )
task = self._get_cached_task(task_hash_legacy) or self._get_cached_task(task_hash) task = self._get_cached_task(task_hash)
# if we found a task, just use # if we found a task, just use
if task: if task:
if disable_clone_task and self.task and self.task.status == self.task.TaskStatusEnum.created: if disable_clone_task and self.task and self.task.status == self.task.TaskStatusEnum.created: