When the docker context is large the processing looks like it is hanging just before training begins, because the entire context has to be copied to docker.
We should print a warning before mflow.run if the docker context directory is over a certain threshold so that the user is aware it might take a while and to prompt them to delete unnecessary/temp folders included by mistake.
When the docker context is large the processing looks like it is hanging just before training begins, because the entire context has to be copied to docker.
We should print a warning before mflow.run if the docker context directory is over a certain threshold so that the user is aware it might take a while and to prompt them to delete unnecessary/temp folders included by mistake.