In CodaLab, we have the possibility to add data directly inside the compute workers, so it can be read by the ingestion or scoring program:
https://github.com/codalab/codalab-competitions/wiki/User_Using-your-own-compute-workers#put-data-directly-inside-the-worker
The way it is implemented is simply that the folder /data of the worker is mounted on the docker container of the submission.
Do we have this on Codabench? If no, should we implement this in the same way than in CodaLab or is there a better practice?
In CodaLab, we have the possibility to add data directly inside the compute workers, so it can be read by the ingestion or scoring program:
https://github.com/codalab/codalab-competitions/wiki/User_Using-your-own-compute-workers#put-data-directly-inside-the-worker
The way it is implemented is simply that the folder
/dataof the worker is mounted on the docker container of the submission.Do we have this on Codabench? If no, should we implement this in the same way than in CodaLab or is there a better practice?