mirror of
https://github.com/hpcaitech/ColossalAI.git
synced 2025-05-05 23:18:22 +00:00
7080a8edb0
* [workflows]bug_repair * [workflow]new_pr_fixing_bugs Co-authored-by: binmakeswell <binmakeswell@gmail.com>
17 lines
665 B
Bash
17 lines
665 B
Bash
pip install -r requirements.txt
|
|
|
|
# distplan in ["colossalai", "zero1", "zero2", "torch_ddp", "torch_zero"]
|
|
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
|
|
export BATCH_SIZE=8
|
|
export MODEL_TYPE="gpt2_medium"
|
|
|
|
|
|
mkdir -p logs
|
|
torchrun --standalone --nproc_per_node=${GPUNUM} train_gpt_demo.py --tp_degree=${TPDEGREE} --model_type=${MODEL_TYPE} --batch_size=${BATCH_SIZE} --placement ${PLACEMENT} --shardinit ${USE_SHARD_INIT} --distplan ${DISTPAN} 2>&1 | tee ./logs/${MODEL_TYPE}_${DISTPAN}_gpu_${GPUNUM}_bs_${BATCH_SIZE}_tp_${TPDEGREE}.log
|