Is Task.current_task() creating a task?
Hmm it should not, it should return a Task instance if one was already created.
That said, I remember there was a bug (not sure if it was in a released version or an RC) that caused it to create a new Task if there isn't an existing one. Could that be the case ?
Update us if it solved the issue (for increased visibility)
Hi SpotlessLeopard9
I got many tasks that were just hang at the end of the script without ...
I remember this exact issue was fixed with 1.1.5rc0, see here:
https://clearml.slack.com/archives/CTK20V944/p1634910855059900
Can you verify with the latest RC?pip install clearml==1.1.5rc3
CheerfulGorilla72 sounds like a great idea, I'll pass along to documentation ppl 🙂
data it is going to s3 as well as ebs. Why so it should only go to s3
This sounds odd, if this is mounted then it goes to the S3 (the link will point to the files server, but it will be stored on the mounted drive i.e. S3)
wdyt?
You mean parameters of the pipeline? Is this a pipeline from Tasks or from function decorator?
HungryArcticwolf62 the new clearml-serving is almost out (eta late next week), you can already start playing here:
https://github.com/allegroai/clearml-serving/tree/dev
Example:
train+serve
https://github.com/allegroai/clearml-serving/tree/dev/examples/sklearn
okay that makes sense, if this is the case I would just use clearml-agent execute --id <task_id here>
to continue the training Task.
Do notice you have to reload your last chekcpoint from the Task's models/artifacts to continue 🙂
Last question, what is the HPO optimization algorithm, is it just grid/random search or optuna hbop/optuna, if this is the later, how do make it "continue" ?
Why would that require refactoring ? Dataset class should take care if it internally ,no?
The reason my_name is a subproject , is that so every version could be a "Task" inside that project , just easier to manage (or at least that was the idea)
could you send the entire log here?
i.e. from the "docker-compose" command line and onward
Hi RoundMosquito25
This is a bit old but probably a good start:
https://clear.ml/blog/stacking-up-against-the-competition/
tl;dr
ClearML advantages (at least a few I can think of)
Scales way better Enables out of the box experiment orchestration (i.e. remote execution etc) Data management Nicer UI Full RestAPI Full MLops platform Model serving Query-able model repositoryProbably more 🙂
Docker cmd is basically docker image name but you can add parameters as well.
For example "Nvidia/cuda" or "Nvidia/cuda -v /mnt/data:/mnt/data"
What do you mean? every Model has a unique ID, what do you consider a version?
Ohh! I see now
@<1526371965655322624:profile|NuttyCamel41> the "backend: "pytorch" is not really supported because it does not use the optimized Triron engine (which is the reason to run Triron server)
In order to use pytorch you need to convert it to torchscript and then deploy, see example here:
None
[None](https://github.com/allegroai/clearml-serving/blob/7ba356efc97a6ae2159283d198d981b3c1ab85e6/examples/pytor...
RoundMosquito25 do notice the agent is pulling the code from the remote repo, so you do need to push the local commits, but the uncommitted changes clearml will do for you. Make sense?
Hi HappyDove3task.set_script
is a great way to add the info (assuming the .git is missing)
Are you running it using PyCharm? (If so use the clearml pycharm plugin, it basically passes the info from your local git to the remote machine via OS environment)
It runs directly but leads to the above error with clearml
Both manually (i.e. calling Task.init and running it without agent, and with agent ? same exact behavior ?
You should manually remove the cudatoolkit from the installed packages section in the UI, then try to send it to the agent and see if it works. The question is how it ended there in the first place
SweetGiraffe8 Task.init will autolog everything (git/python packages/console etc), for your existing process.
Task.create purely creates a new Task in the system, and lets' you manually fill in all the details on that Task
Make sense ?
The problem is not really for the agents to wait (this is easily solved by additional high priority queue) the problem is will you have a "free" agent... you see my point ?
Building the pipeline in runtime from external configuration is very cool!!
I think nested components is exactly the correct solution, and it is a great use case.
I see now.
Let's assume you know which snapshot that was:
` prev_task = Task.get_task(task_id='the_first_training_task_id')
get the second from last checkpoint
task.models['output'][-2].url
prev_scalars = prev_task.get_reported_scalars()
new_task = Task.init('example', 'new task')
logger = new_task.get_logger()
do some fpr loop and report the prev_scalars with logger.report_scalars
new_task.flush(wait_for_uploads=True)
new_task.set_initial_iteration(22000)
start the train `
task = Task.init(...) if task.running_locally(): # wait for the repo detection and requirements update task._wait_for_repo_detection() # reset requirements task._update_requirements(None)
🙂
Hi NastyFox63
This seems like most of the reports are converted to pngs (which is what the automagic does if it fails to convert the matplotlib into interactive plot).
no more than 114 plots are shown in the plots tab.
Are you saying we have 114 limit on plots ?
Is this true for "full screen" mode (i.e. not in the experiments table but switch to full detailed view)
Hi PunyGoose16 ,
I think the website is probably the easiest 🙂
https://clear.ml/contact-us/
I think they get back to quite quickly
@<1545216070686609408:profile|EnthusiasticCow4>
Is there currently a way to bind the same GPU to multiple queues? I believe the agent complains last time I tried (which was a bit ago)
run multiple agents on the same GPU,
CLEARML_WORKER_NAME=host-gpu0a clearml-agent daemon --gpus 0
CLEARML_WORKER_NAME=host-gpu0b clearml-agent daemon --gpus 0
one can containerise the whole pipeline and run it pretty much anywhere.
Does that mean the entire pipeline will be running on the instance spinning the container ?
From here: this is what I understand:
https://kedro.readthedocs.io/en/stable/10_deployment/06_kubeflow.html
My thinking was I can use one command and run all steps locally while still registering all "nodes/functions/inputs/outputs etc" with clearml such that I could also then later go into the interface and clone an...
GrumpyPenguin23 could you help and point us to an overview/getting-started video?
Sorry if it's something trivial. I recently started working with ClearML.
No worries, this has actually more to do with how you work with Dask
The Task ID is the unique id of the any Task in the system (task.id will return the UID str)
Can you post a toy Dash code here, I'll explain how to make it compatible with clearml 🙂