![Profile picture](https://clearml-web-assets.s3.amazonaws.com/scoold/avatars/SarcasticSparrow10.png)
Reputation
Badges 1
58 × Eureka!I come across many small questions like these which may been answered earlier. But they are hard to find in Slack messages. Is it better to post such questions on Stackoverflow so they benefit everybody? I might post the link here.
(Do notice that even though you can spin two agents on the same GPU, the nvidia drivers cannot share allocated GPU memory, so if one Task consumes too much memory the other will not have enough free GPU memory to run)
Basically the same restriction as manually launching two processes using the same GPU
That makes sense. Currently, I use python multiprocessing to launch multiple experiments on the sam GPU device. I'm guessing using trains-agent
will be similar
Got it. I haven't tried setting up trains-agent
yet so I don't know much about the overhead of launching the agent. I'd imagine if it has to create the full environment (installing requirements, etc), the overhead might not be that low. But as I'm reading, it looks like I can use a docker image with the full env. Is my understanding correct?
You will need to habe multiple
trains-agent
s but they will be sharing the same queue (i.e. pulling jobs from the same queue the HPO process is pushing to)
Make sense ?
Hmm. So say I have a parameter NUM_PARALLEL_EXECUTIONS
, I can programmatically launch that many trains-agent
for every optimization run?!
Ok, So Git credentials are present at two locations - 1) outside the agent
config and 2) inside it. I updated credentials at both locations and now I'm seeing agent.git_user = <username>
in the dump, but I still have the same issue.
` # Set GIT user/pass credentials
leave blank for GIT SSH credentials ...
2. interesting error, maybe we can revert to "thread mode" if running under a daemon. (I have to admit, I'm not sure why python has this limitation, let me check it...)
Yes, I'm not sure either. I have banged my head against the wall in trying to have multiple level of subprocesses, but it gets too complicated with python. Let me know what you find out
Yes, I am using Pool. Here is what I think is happening. clearml launches a subprocess which I assume is a daemonic process. That process in-turn launches a subprocess for training which causes the error I mentioned
Yes the 'training' is my main code. You can think of it has launching a job (training or inference). My main code launches multiple jobs using multiprocessing. Each job is a seprate task for clearml that gets logged. Does that make sense?
The second subprocess is by design. It becomes the primary process when clearml does not use multiprocessing. I hope I'm not confusing you further
fatal: could not read Username for '
': terminal prompts disabled error: Could not fetch origin
Why is trains-agent trying read from terminal prompt instead of trains.conf
?
I'm using docker to run the experiment. Could it be that the config in the docker container doesn't have the git credentials?
SuccessfulKoala55
For security reasons I don't want to have my password written out in a file. I'm trying to use https://docs.github.com/en/free-pro-team@latest/github/authenticating-to-github/creating-a-personal-access-token (PAT) from Github but I get authentication error. Is there an issue using PAT?
SuccessfulKoala55 Yes, I am using the --docker flag.
You are right about the Keyring. Once I make sure credentials are stored in a secure way, it works as expected. Thanks :)
That makes sense. The configuration file is located at ~/trains.conf
which I believe is the default location.
No I can't see my username printed out in the dump