hub-er / distribuuuu Goto Github PK

The pure and clear PyTorch Distributed Training Framework.

• Introduction
• Requirements and Usage
  • Dependency
  • Dataset
  • Basic Usage
  • Slurm Cluster Usage
• Baselines
• Zombie processes problem
• Acknowledgments
• Citation

Distribuuuu is a Distributed Classification Training Framework powered by native PyTorch.

Please check tutorial for detailed Distributed Training tutorials:

• Single Node Single GPU Card Training [snsc.py]
• Single Node Multi-GPU Crads Training (with DataParallel) [snmc_dp.py]
• Multiple Nodes Multi-GPU Cards Training (with DistributedDataParallel)
  • torch.distributed.launch [mnmc_ddp_launch.py]
  • torch.multiprocessing [mnmc_ddp_mp.py]
  • Slurm Workload Manager [mnmc_ddp_slurm.py]
• ImageNet training example [imagenet.py]

For the complete training framework, please see distribuuuu.

• Install PyTorch>= 1.5 (has been tested on 1.5, 1.7.1 and 1.8)
• Install other dependencies: pip install -r requirements.txt

Download the ImageNet dataset and move validation images to labeled subfolders, using the script valprep.sh.

ILSVRC
|_ train
|  |_ n01440764
|  |_ ...
|  |_ n15075141
|_ val
|  |_ n01440764
|  |_ ...
|  |_ n15075141
|_ ...

mkdir -p /path/to/distribuuuu/data

ln -s /path/to/ILSVRC /path/to/distribuuuu/data/ILSVRC

Single Node with one task

# 1 node, 8 GPUs
python -m torch.distributed.launch \
    --nproc_per_node=8 \
    --nnodes=1 \
    --node_rank=0 \
    --master_addr=localhost \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml

Distribuuuu use yacs, a elegant and lightweight package to define and manage system configurations. You can setup config via a yaml file, and overwrite by other opts. If the yaml is not provided, the default configuration file will be used, please check distribuuuu/config.py.

python -m torch.distributed.launch \
    --nproc_per_node=8 \
    --nnodes=1 \
    --node_rank=0 \
    --master_addr=localhost \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml \
    OUT_DIR /tmp \
    MODEL.SYNCBN True \
    TRAIN.BATCH_SIZE 256

# --cfg config/resnet18.yaml parse config from file
# OUT_DIR /tmp            overwrite OUT_DIR
# MODEL.SYNCBN True       overwrite MODEL.SYNCBN
# TRAIN.BATCH_SIZE 256    overwrite TRAIN.BATCH_SIZE

# 1 node, 2 task, 4 GPUs per task (8GPUs)
# task 1:
CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch \
    --nproc_per_node=4 \
    --nnodes=2 \
    --node_rank=0 \
    --master_addr=localhost \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml

# task 2:
CUDA_VISIBLE_DEVICES=4,5,6,7 python -m torch.distributed.launch \
    --nproc_per_node=4 \
    --nnodes=2 \
    --node_rank=1 \
    --master_addr=localhost \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml

# 2 node, 8 GPUs per node (16GPUs)
# node 1:
python -m torch.distributed.launch \
    --nproc_per_node=8 \
    --nnodes=2 \
    --node_rank=0 \
    --master_addr="10.198.189.10" \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml

# node 2:
python -m torch.distributed.launch \
    --nproc_per_node=8 \
    --nnodes=2 \
    --node_rank=1 \
    --master_addr="10.198.189.10" \
    --master_port=29500 \
    train_net.py --cfg config/resnet18.yaml

# see srun --help 
# and https://slurm.schedmd.com/ for details

# example: 64 GPUs
# batch size = 64 * 128 = 8192
# itertaion = 128k / 8192 = 156 
# lr = 64 * 0.1 = 6.4

srun --partition=openai-a100 \
     -n 64 \
     --gres=gpu:8 \
     --ntasks-per-node=8 \
     --job-name=Distribuuuu \
     python -u train_net.py --cfg config/resnet18.yaml \
     TRAIN.BATCH_SIZE 128 \
     OUT_DIR ./resnet18_8192bs \
     OPTIM.BASE_LR 6.4

Baseline models trained by Distribuuuu:

Before PyTorch1.8, torch.distributed.launch will leave some zombie processes after using Ctrl + C, try to use the following cmd to kill the zombie processes. (fairseq/issues/487):

kill $(ps aux | grep YOUR_SCRIPT.py | grep -v grep | awk '{print $2}')

PyTorch1.8 is suggested, which fixed the issue about zombie process. (pytorch/pull/49305)

Provided codes were adapted from:

• facebookresearch/pycls
• pytorch/examples
• open-mmlab/mmcv

I strongly recommend you to choose pycls, a brilliant image classification codebase and adopted by a number of projects at Facebook AI Research.

@misc{bigballon2021distribuuuu,
  author = {Wei Li},
  title = {Distribuuuu: The pure and clear PyTorch Distributed Training Framework},
  howpublished = {\url{https://github.com/BIGBALLON/distribuuuu}},
  year = {2021}
}

Feel free to contact me if you have any suggestions or questions, issues are welcome, create a PR if you find any bugs or you want to contribute. 🍰