Reputation
Badges 1
533 × Eureka!There are many ohter packages in my environment which are not listed
I don't htink I can, this is private IP and to create a dummy example of a pipeline and execution will take me more time than I can dedicate to this
I believe that is why MetaFlow chose conda
as their package manager, because it can take care of these kind of dependencies (even though I hate conda 😄 )
I guess not many tensorflowers running agents around here if this wasn't brought up already
For example I have a DATA_DIR
environment variable which points to the directory where disk-data is stored
but shouldn't the :lastest
make it redownload the right image?
when I specify --packages
I shoudl manually list them all not?
UptightCoyote42 - How are these images avaialble to all agents? Do you host them on Docker hub?
cool, didn't know about the PAT
Wait but I don't want to execute it
` # define pipeline
pipe = clearml.PipelineController(
name=TASK_NAME,
project=PROJECT_NAME,
version='0.0.1',
add_pipeline_tags=False,
)
pipe.set_default_execution_queue('default')
Adding steps
pipe.add_step(name=f'{start_date_train}_{end_date_train}_choose_best',
base_task_project=CHOOSE_PROJECT_NAME,
base_task_name=CHOOSE_TASK_NAME,
parameter_override=params_override,
...
Okay so at the first part of the code, we define some kind of callback that we add to our steps, so later we can collect them and attach the results to the pipeline task. It looks something like this
` class MedianPredictionCollector:
_tasks_to_collect = list()
@classmethod
def collect_description_tables(cls, pipeline: clearml.PipelineController, node: clearml.PipelineController.Node):
# Collect tasks
cls._tasks_to_collect.append(node.executed)
@classmethod...
how do I run this wizard? is this wizard train's or aws's?
Is tehre anything specific about the logs we're looking for? Because if I just dumop them it will take me a while to see no sensitive data and naming is there
so basically - if she has new commits locally that werent pushed it won't work
But if she did not commit her latest changes, and now she enqueues - it will work?
If this includes scheduling through pipelines, in my opinion there should be an option to execute a pipeline without an agent. Sometimes for development I just want to execute a pipeline on my local machine just as I would a task...
I mean usually it would read if cached_file: return cached_file