Reputation
Badges 1
25 × Eureka!Also, how do pipelines compare here?
Pipelines are a type of Task, so like Tasks you can clone and enqueue them, or set them as the target of the trigger.
the most flexible solution would be to have some way of triggering the execution of a script in the parent task environment,
This is the exact idea of the TriggerScheduler None
What am I missing here?
Hi @<1523701260895653888:profile|QuaintJellyfish58>
Is there a way or a trigger to detect when the number of workers in a queue reaches zero?
You mean to spin them down? what's the rational ?
Iād like to implement a notification system that alerts me when there are no workers left in the queue.
How are they "dropping" ?
Specifically to your question, let me check I'm sure there is an API that get's that data becuase you can see it in the UI š
Hi VexedCat68
One of my steps just finds the latest model to use. I want the task to output the id, and the next step to use it. How would I go about doing this?
When you say "I want the task to output the id" do you mean to pass t to the next step:
Something like this one:
https://github.com/allegroai/clearml/blob/c226a748066daa3c62eddc6e378fa6f5bae879a1/clearml/automation/controller.py#L224
but I can't seem to figure out a way to do something similar using a task in add_step
VexedCat68 With "add_step" it assumes the Task you are adding is self contained (i.e. there is no "return object" to serialize), this means you can only add arguments, or use the artifacts the Task (i.e. step) will recreate, obviously you knowing in advance what the step creates. Make sense ?
VexedCat68 yes š you can also pass the parent folder and it will zip the entire subfolders into a single artifact
I can read them programmatically using tensorboard and the log the using clearml logger,
StaleButterfly40 this will be a great script to put somewhere (I'm sure you are not the only one with this problem). Maybe put it as a GitHub issue ? wdyt ?
This is by design, they cannot use the exact same venv because if the code starts creating files/change them it happens inside the venv and might cause them to crash.
That said if you are running with venv cache, the first one will create the venv and the second one will create a copy from the cache.
Hi @<1523701260895653888:profile|QuaintJellyfish58>
Based on the docs
None
I think this should have worked, are you running the actual task_scheduler
on yout machine? on the services queue ? what's the console output you see there ?
VexedCat68
. So the checkpoints just added up. I've stopped the training for now. I need to delete all of those checkpoints before I start training again.
Are you uploading the checkpoints manually with artifacts? or is it autologged & uploaded ?
Also why no reuse and overwrite older checkpoints ?
Hi! I was wondering why ClearML recognize Scikit-learn scalers as Input Models...
Hi GiganticTurtle0
any joblib.load/save is logged by clearml (it cannot actually differentiate what it is used for ...)
You can of course disable it with Task.init(..., auto_connect_frameworks={'joblib': False})
Sadly, I think we need to add another option like task_init_kwargs
to the component decorator.
what do you think would make sense ?
this sounds like docker build issue on macos M1
https://pythonspeed.com/articles/docker-build-problems-mac/
Hi WackyRabbit7
the services
(or the agent running there) is spinning multiple Tasks (as opposed to regular agent where it is one task at a time).
how can I give this agent git access?
in the docker-compose you can configure the git credentials (user/pass or user/key it is the same).
https://github.com/allegroai/clearml-server/blob/d0e2313a24eb1248ebf0ddf31bf589de0d675562/docker/docker-compose.yml#L137
Hmm can you try with additional configuration, next to "secure: true" in your clearml.conf, can you add "verify: false"
I can but that is not a configuration we would want to run with in production
Agreed, I just want to isolate the issue. I think this is the bottom python interface missing some configuration or environment variables
so that you can get the latest artifacts of that experiment
what do you mean by " the latest artifacts "? do you have multiple artifacts on the same Task or s it the latest Task holding a specific artifact?
If possible, can we have a "only one experiment can be given a single tag"
You mean "moving a tag" automatically (i.e. if someone else had the same tag it is removed from it)?
Yes, the webserver doesn't know where the api server is, it will access /api and then the nginx running the webapp will do the routing (reverse proxy)
I think that for some reason it is failing to do that (actually similarly to the stackoverflow you linked)
Hi ShakyJellyfish91
It seems clearml is using a single connection, that takes a long time download
Hmm, I found this one:
https://github.com/allegroai/clearml/blob/1cb5dbb276026644ae20fef63d58256cdc887818/clearml/storage/helper.py#L1763
Does max_connections=10
mean 10 concurrent connections ?
The "Optimizer task" will continue to run as long as there are sub-Tasks it launched.
Is anything else running/pending ?
It completed after the max_job limit (10)
Yep this is optuna "testing the water"
Yeah, Curious - is a lot of clearml usecases not geared for notebooks?
That is somewhat correct, notebooks are not actually used with a lot of deep-learning projects as they require entire repository to support.
I guess generally speaking the workflow is, "test your code" (i.e. small scale with limited data), then clone and enqueue for remote execution.
That said, I think it will be great to expand the support.
TrickySheep9 I like the idea of context for Tasks, can you expand on how...
Hi LazyTurkey38
Documentation for applications is currently worked on, generally speaking this is a way to package features available in ClearML with a UI interface. First these are going to be applications built by the ClearML team and later expanded for the community to be able to contribute to them. Finally users will be able to add their own applications (i.e. package Tasks with UI wizard and dashboard) in their hosted solutions. wdyt?
Hmm do you host it somewhere? Is it pre-installed on the container?
I have a client that runs clearml-session and i saw from the agent's logs that the installation of vscode fails.
That makes sense, it downloads the vscode in runtime, do you have an alternative location? or maybe it is easier to built a container with the vscode pre installed ?
Could not install packages due to an EnvironmentError: [Errno 2] No such file or directory: '/tmp/build/80754af9/attrs_1604765588209/work'
Seems like pip failed creating a folder
Could it be you are out of space ?
SubstantialElk6
Regrading cloning the executed Task:
In the pip requirements syntax, "@" is a hint that tells pip where to find the package if it is not preinstalled.
Usually when you find the @ /tmp/folder
It means the packages was preinstalled (usually pre installed in the docker).
What is the exact scenario that caused it to appear (this was always the case, before v1 as well).
For example zipp
package is installed from pypi be default and not from local temp file.
Your fix b...