2022-12-20 06:30:27 +00:00
|
|
|
# distplan in ["colossalai", "zero1", "zero2", "torch_ddp", "torch_zero"]
|
2022-11-16 03:36:27 +00:00
|
|
|
export DISTPAN="colossalai"
|
|
|
|
|
|
|
|
|
|
# The following options only valid when DISTPAN="colossalai"
|
|
|
|
|
export TPDEGREE=2
|
|
|
|
|
export GPUNUM=4
|
|
|
|
|
export PLACEMENT='cpu'
|
|
|
|
|
export USE_SHARD_INIT=False
|
|
|
|
|
|
|
|
|
|
env OMP_NUM_THREADS=16 torchrun --standalone --nproc_per_node=${GPUNUM} train_gpt_demo.py --tp_degree=${TPDEGREE} --placement ${PLACEMENT} --shardinit ${USE_SHARD_INIT} --distplan ${DISTPAN} 2>&1 | tee run.log
|
