EnormousCormorant39 , there are SDK methods for using the datasets. I think this will simplify your process immensely.
https://clear.ml/docs/latest/docs/references/sdk/dataset
Also here is a small example for the usage 🙂
` task = Task.init(project_name="<PROJECT_NAME>", task_name="<TASK_NAME>")
#Create dataset
ds = Dataset.create(dataset_name="<DATASET_NAME>", dataset_project="PROJECT_NAME")
ds.add_files("<PATH_TO_FILE/S>")
ds.upload()
ds.finalize() `
Yes. CostlyOstrich36
I’m running on an on-prem machine, and trying to automate the whole process of training. meaning bringing the data -> creating the dataset -> running the training task.
When creating a dataset and adding files it seems there is a Task being created in the background (which is awesome).
My question is - can I run this task on a remote machine instead of the machine which i’m running the CLI on?
Hi EnormousCormorant39 ,
is there a way to enqueue the dataset
add
command on a worker
Can you please elaborate a bit on this? Do you want to create some sort of trigger action to add files to a dataset?