ClearML - Auto-Magical CI/CD to streamline your AI workload. Experiment Management, Data Management, Pipeline, Orchestration, Scheduling & Serving in one MLOps/LLMOps solution
Task.get_requirements()
method returning the task’s requirementsStorageManager.download_folder()
using the max_workers
argumentTask.init()
force_requirements_env_freeze
Task.set_packages()
docstring notesOutputModel
while in offline mode results in an error (#1172)gpu_memory
and gpu_type
is not being collected in some casesclearml-data
CLI tool will move non-dataset tasks to a .dataset
project if the respective task is not a datasetDataFrame
artifacts with hierarchical indices get mangled by the CSV round-tripurllib3
sends deprecation warning when setting ssl_version
Task.connect(dict)
return value is not dict-compatiblejsonargparse
sub-command config parsing.
in its namesecure
property not forwarded in StorageHelper
(#1205, thanks @odulcy-mindee!)NVIDIA_VISIBLE_DEVICES=all
env var value will result in no GPU metrics being reportedoutput_uri
when an AWS S3 config section with secure=false
exists raises an errorCacheManager
attempts to clear cache when file limit has not been reached resulting in slow unnecessary checkstqdm
if installedPipelineController.version
propertyignore_remote_overrides
argument to Task.connect()
and Task.connect_configuration()
allowing users to disable overriding values when running remotelyCLEARML_API_HOST_VERIFY_CERT
environment variable (instead of just allowing the user to disable verification)tempfile.mktemp
distutils
TaskScheduler
does not save local state before running remotelyPipelineController.abort_on_failure
is set to False and a step has been abortedcontinue_last_task=True
torch.distributed
behavior and auto-detect Huggingface accelerate
distributed executionjsonargparse
integration: allow_config_file_override_from_ui
argument added for enabling/disabling overriding configuration file entriesraise_on_error
parameter to Model.update_weights()
and Model.update_weights_package()
HyperParameterOptimizer
classtask.set_repo()
tensorflow-intel
as tensorflow
langchain
integration (#1124, #1126)target_project
with TaskScheduler.add_task()
corrupts project if used with pipelines (#1137)PipelineController.add_function_step()
fails when adding custom decorated step (#1154)jsonargparse
binding does not capture parameters before Task.init()
is called (#1164)jsonargparse
arguments should be deletedTask.set_repo()
and Task.set_packages()
do not wait for automatic asynchronous repository detectionBackgroundMonitor
lightgbm
model loading from a string (#1136, thanks @bedapisl!)omegaconf
output_uri
in pipeline decorators and functionsjsonargparse
argumentsdefault_output_uri
on remote runs when Task.init(output_uri=False)
was specified and the backend task data has no output_uri
specifiedLoginError
when initializingDeprecationWarning
in newer versions (#1128, thanks @xquyvu!)task.set_script(diff="")
does not clear the uncommitted changes (#1106)attrs.__version__
deprecation warning (#1145)Task.get_parameters_as_dict(cast=True)
raises a YAML parser errorself.lock
to None
Model.connect()
in remote execution might result in the wrong model being connectedConnectionError
(which actually indicates that a read has timed out, and should be retried)pyjwt
requirements for Python < 2.9controller.add step()
parameter overrides (#1099, thanks @natephysics!)force_single_script_file
argument to Task.create()
to avoid git repository auto detectionos.register_at_fork
instead of monkey patching fork for python > 3.6Model.archive()
and Mode.unarchive()
methods (#1096)Exception [Errno 17] File exists
warning)sys
should not appear in pipeline component auto detected packages