capreolus-ir / capreolus

A toolkit for end-to-end neural ad hoc retrieval
https://capreolus.ai
Apache License 2.0
95 stars 32 forks source link

monoBERT resource allocation #172

Closed d1shs0ap closed 3 years ago

d1shs0ap commented 3 years ago

While trying to train monoBERT, I tried to enter a SLURM job with the following line: srun --mem=48G --cpus-per-task=2 --time=8:0:0 --gres=gpu:v100l:4 --pty zsh

However, I receive the error message: srun: error: Unable to allocate resources: Requested time limit is invalid (missing or exceeds some limit)

Is my SLURM job submission correct? Am I allocating too much (or too little) resources? If not, then how should I resolve this error?

Thanks in advance!

crystina-z commented 3 years ago

@d1shs0ap heyy sorry I somehow thought I replied to this.

cedar allows up to 3 hours for srun request. For a job with longer running time we can switch to sbatch, an example script can be found here.

d1shs0ap commented 3 years ago

@crystina-z Great, thanks!