Hi UnsightlySeagull42 , you can always use a relative or absolute path (that's just Python, not related to ClearML). What exactly do you need to access?

I need to access to a tfrecord file
but with allegro i do not have access to the folder

Are you running the task locally? Or using a ClearML agent?

first i run it locally. This works. But then I use the clearML agent and then it does not work

i do not have access to the folder

What does that mean? Are you getting an error?

assert os.path.exists("path")
with this line I get the error that the path does not exist

quick question if I may, are you running clearml-agent with --docker mode or without ? are you running the clearml-agent on the same machine as the path exists on another machine entirely ?

i run it with docker mode

i run it on the same machine

understand. Are you comfortable with docker ? If so, I would probably suggest doing a docker run -it <identifier> bash and seeing if that folder does, indeed, exist in the docker image

Sorry.. we are currently swamped with ODSC East requests and presentations etc 😄

UnsightlySeagull42 as AlertBlackbird30 kind of pointed out, the issue is probably that when you're running your agent in docker mode, it runs the task inside the docker container which has its own virtual file system (so your external folder/file is not found there). In order for your folder/file to be accessible, you'll need to add a volume mount - we'll be happy to explain when you have the time 🙂

thanks for the info
I have time now
how can I use "volume mount" with allegro?

What exactly are you running (agent configuration)?

clearml-agent --config-file /home/chuber/clearml.conf daemon --detached --gpus 1
--queue KA_ML2_GPU1 --docker nvidia/cuda:10.1-cudnn7-devel

And is the external path you want to mount also found in the machine where the agent is running?

if i start the script via the console then the path is available but if i start a training via the graphical interface of allegro then not

i think there is a problem with docker. because i am in the docker container and I have to use volume mount to get access to paths outside the container.
But i dont know how to use volume mount exactly

i try to run the agent without docker. Without docker mode the path is available. But i need docker for tensorflow and cuda