Alright. Thanks CostlyOstrich36

Yup, was thinking of bash script.

The intent is to generate some outputs from the clearml task and thinking probably to package it into a docker image for ease of sharing to others that are not plug into our network and able to run the image directly.

I guess we need to understand the purpose of the various states. So far only "archive, draft, publish". Did I miss any?

Yes. The training working well with cuda

May I know where to set the cert to in env variable?

seems like it was broken for numpy version 1.24.1.
Tried with numpy 1.23.5 and it works.

Hi Bart, yes. Running with inference container.

https://clear.ml/docs/latest/docs/integrations/storage/
Try add the <path to your cert> for s3.credentials.verify.

I see. Was wondering any advantage to do it any of the ways.

Example i build my docker image using a image in docker hub. In this image, i installed torch and cupy packages. But when i run my experiment in this image, the packages are not found.

Yes, I ran the experiment inside.

@<1526734383564722176:profile|BoredBat47> Just to check if u need to do update-ca-certificates or equivalent?

SdK meaning I run the agent using clearml-agent daemon ....
Alternatively I understand I can also run the agent using docker run allegroai/clearml-agent:latest. But I cannot figure out how to add --restart, --queue, -- gpus flag to the container

By the way, how can I start up the clearml agent using the clearml-agent image instead of SDK? Do u have an example of the docker run command that includes the queue, gpus etc?

@<1523701070390366208:profile|CostlyOstrich36> This is output_uri or where do I put this url?

Cool thanks guys. I am clearer now. Was confused by the obsolete info. Thanks for the clarification.

@<1523701205467926528:profile|AgitatedDove14> do u mean not using helm but fill up the values and install with the yaml files directly? E.g. kubectl apply ...

SuccessfulKoala55 i tried comment off fileserver, clearml dockers started but it doesn't seems to be able to start well. When I access clearml via webbrowser, site cannot be reached.

Just to confirm, I commented off these in docker-compose.yaml.

apiserver:
command:
- apiserver
container_name: clearml-apiserver
image: allegroai/clearml:latest
restart: unless-stopped
volumes:
- /opt/clearml/logs:/var/log/clearml
`...

Not exactly sure yet but I would think user tag for deployed make sense as it should be a deliberated user action. And additional system state is required too since a deployed state should have some pre-requitise system state.

I would also like to ask if clearml has different states for a task, model, or even different task types? Right now I dun see differences, is this a deliberated design?

Ah I think I was not very clear on my requirement. I was looking at porting project level, not entire clearml data over. Is it possible instead?

Yes. But I not sure what's the agent running. I only know how to stop it if I have the agent id

Nice. It is actually dataset.id .

Is there a way to troubleshoot further?

Thanks AgitatedDove14 and TimelyMouse69 . The intention was to have some traceability between the two setups. I think the best way is to enforce some naming convention (for project and name) so we can know how they are related? Any better suggestions?

When I run as regular remote task it works. But when I run as a step in pipeline, it cannot access the same folder in my local machine.

It return false. Just to share abit more, I have the requirements.txt in gitlab with my codes and are in folders. Do I need to provide a gitlab path?

@<1523701205467926528:profile|AgitatedDove14> when my codes get the clearml datasets, it stores in the cache e.g. /$HOME/.clearml/cache....

I wanted it to be in a mounted PV instead, so other pods (in same node) who needed same datasets can use without pulling again.

OK let me try by adding to vol mount.

Thanks I just realised I didn't add --docker

@<1523701070390366208:profile|CostlyOstrich36> Yes. I'm running on k8s