Closed babla9 closed 5 months ago
Hi, I'm trying to run this finetuning SPHINX script on a 8xV100 machine.
However, my machine doesnt have SLURM installed, so how do I allow the script to access all gpus on the cluster without using srun / slurm?
Thanks!
replace srun python xxx with torchrun --nproc_per_node=8 xxx should be all you need
srun python xxx
torchrun --nproc_per_node=8 xxx
Hi, I'm trying to run this finetuning SPHINX script on a 8xV100 machine.
However, my machine doesnt have SLURM installed, so how do I allow the script to access all gpus on the cluster without using srun / slurm?
Thanks!