RuntimeError: unable to mmap 276 bytes from file </torch_116128_1191909093_62920>: Cannot allocate memory (12)

[ahmadjaved97]

While trying to run the training, whenever I put the NUM_GPU > 1, it shows the following error

. I am using the following slurm command to allocate the resources srun --pty --nodes=1 --ntasks-per-node=8 --cpus-per-task=8 --time=7-00:00:00 --partition=gpu_partition --gpus=4 /bin/bash. I am trying to train a multilabel dataset