拉起多卡训练脚本示例

构建模型脚本

# 导入依赖和库 
import torch 
from torch import nn 
import torch_npu 
import torch.distributed as dist 
from torch.utils.data import DataLoader 
from torchvision import datasets 
from torchvision.transforms import ToTensor 
import time 
import torch.multiprocessing as mp 
import os 
 
torch.manual_seed(0) 
# 下载训练数据 
training_data = datasets.FashionMNIST( 
    root="./data", 
    train=True, 
    download=True, 
    transform=ToTensor(), 
) 
 
# 下载测试数据 
test_data = datasets.FashionMNIST( 
    root="./data", 
    train=False, 
    download=True, 
    transform=ToTensor(), 
) 
 
# 构建模型 
class NeuralNetwork(nn.Module): 
    def __init__(self): 
        super().__init__() 
        self.flatten = nn.Flatten() 
        self.linear_relu_stack = nn.Sequential( 
            nn.Linear(28*28, 512), 
            nn.ReLU(), 
            nn.Linear(512, 512), 
            nn.ReLU(), 
            nn.Linear(512, 10) 
        ) 
 
    def forward(self, x): 
        x = self.flatten(x) 
        logits = self.linear_relu_stack(x) 
        return logits 
 
def test(dataloader, model, loss_fn): 
    size = len(dataloader.dataset) 
    num_batches = len(dataloader) 
    model.eval() 
    test_loss, correct = 0, 0 
    with torch.no_grad(): 
        for X, y in dataloader: 
            X, y = X.to(device), y.to(device) 
            pred = model(X) 
            test_loss += loss_fn(pred, y).item() 
            correct += (pred.argmax(1) == y).type(torch.float).sum().item() 
    test_loss /= num_batches 
    correct /= size 
    print(f"Test Error: \n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \n")

获取分布式超参数

在模型脚本中，构建主函数main，在其中获取分布式训练所需的超参数。

shell脚本/torchrun/torch_npu_run方式

def main(world_size: int,  batch_size = 64, total_epochs = 5,):    # 用户可自行设置
    ngpus_per_node = world_size
    main_worker(args.gpu, ngpus_per_node, args)

mp.spawn方式

def main(world_size: int,  batch_size = 64, total_epochs = 5,):    # 用户可自行设置
    ngpus_per_node = world_size
    mp.spawn(main_worker, nprocs=ngpus_per_node, args=(ngpus_per_node, args))    # mp.spawn方式启动

Python方式

def main(world_size: int,  batch_size, args):    # 使用Python拉起命令中设置的超参数
    ngpus_per_node = world_size
    args.gpu = args.local_rank    # 任务拉起后，local_rank自动获得device号
    main_worker(args.gpu, ngpus_per_node, args)

设置地址和端口号

在模型脚本中设置地址与端口号，用于拉起分布式训练。由于昇腾AI处理器初始化进程组时initmethod只支持env:// （即环境变量初始化方式），所以在初始化前需要根据实际情况配置MASTER_ADDR、MASTER_PORT等参数。

shell脚本/mp.spawn/torchrun方式的配置代码相同，如下所示：

def ddp_setup(rank, world_size):
    """
    Args:
        rank: Unique identifier of each process
        world_size: Total number of processes
    """
    os.environ["MASTER_ADDR"] = "localhost"    # 用户根据实际情况设置为rank0节点通信ip，单机时可设置为localhost
    os.environ["MASTER_PORT"] = "***"    # 用户根据实际情况设置为rank0节点的端口号
    dist.init_process_group(backend="hccl", rank=rank, world_size=world_size)

torch_npu_run方式，使用时需切换通信方式，修改init_process_group接口的参数。

修改前：init_method使用tcp方式

dist.init_process_group(..., init_method="tcp://xx:**", ......, ) # xx为rank0节点通信ip，**为端口号，根据实际选择一个闲置端口填写

修改后：init_method使用parallel方式

dist.init_process_group(..., init_method="parallel://xx:**", ......, ) # xx为rank0节点通信ip，**为端口号，根据实际选择一个闲置端口填写

Python方式需要把配置参数的命令放到拉起训练中。脚本中代码如下所示：

def ddp_setup(rank, world_size):
    """
    Args:
        rank: Unique identifier of each process
        world_size: Total number of processes
    """
    dist.init_process_group(backend="hccl", rank=rank, world_size=world_size)

添加分布式逻辑

不同的拉起训练方式下，device号的获取方式不同：

shell脚本/torchrun/torch_npu_run方式：在shell脚本中循环传入local_rank变量作为指定的device。
mp.spawn方式：mp.spawn多进程拉起main_worker后，第一个参数GPU自动获得device号（0 ~ ngpus_per_node - 1）。
Python方式：任务拉起后，local_rank自动获得device号。

用户需根据自己选择的方式对代码做不同的修改。下面是修改代码示例：

shell脚本/torchrun/torch_npu_run方式

def main_worker(gpu, ngpus_per_node, args):

    start_epoch = 0
    end_epoch = 5
    args.gpu = int(os.environ['LOCAL_RANK'])    # 在shell脚本中循环传入local_rank变量作为指定的device
    ddp_setup(args.gpu, args.world_size)

    torch_npu.npu.set_device(args.gpu)
    total_batch_size = args.batch_size
    total_workers = ngpus_per_node

    batch_size = int(total_batch_size / ngpus_per_node)    
    workers = int((total_workers + ngpus_per_node - 1) / ngpus_per_node)

    model = NeuralNetwork()

    device = torch.device("npu")

    train_sampler = torch.utils.data.distributed.DistributedSampler(training_data)
    test_sampler = torch.utils.data.distributed.DistributedSampler(test_data)

    train_loader = torch.utils.data.DataLoader(
        training_data, batch_size=batch_size, shuffle=(train_sampler is None),
        num_workers=workers, pin_memory=False, sampler=train_sampler, drop_last=True)

    val_loader = torch.utils.data.DataLoader(
        test_data, batch_size=batch_size, shuffle=(test_sampler is None),
        num_workers=workers, pin_memory=False, sampler=test_sampler, drop_last=True)

    loc = 'npu:{}'.format(args.gpu)
    model = model.to(loc)
    criterion = nn.CrossEntropyLoss().to(loc)
    optimizer = torch.optim.SGD(model.parameters(), lr=1e-3)

    model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu])

    for epoch in range(start_epoch, end_epoch):
        print("curr epoch: ", epoch)
        train_sampler.set_epoch(epoch)
        train(train_loader, model, criterion, optimizer, epoch, args.gpu)


def train(train_loader, model, criterion, optimizer, epoch, gpu):
    size = len(train_loader.dataset)
    model.train()

    end = time.time()
    for i, (images, target) in enumerate(train_loader):
        # measure data loading time

        loc = 'npu:{}'.format(gpu)
        target = target.to(torch.int32)        
        images, target = images.to(loc, non_blocking=False), target.to(loc, non_blocking=False)

        # compute output
        output = model(images)
        loss = criterion(output, target)


        # compute gradient and do SGD step
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        end = time.time()
        if i % 100 == 0:
            loss, current = loss.item(), i * len(target)
            print(f"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]")

mp.spawn方式

不需要专门设置args.gpu，将shell脚本方式中main_worker里的args.gpu均替换为gpu。

def main_worker(gpu, ngpus_per_node, args):

    start_epoch = 0
    end_epoch = 5
    ddp_setup(gpu, args.world_size)

    torch_npu.npu.set_device(gpu)
    total_batch_size = args.batch_size
    total_workers = ngpus_per_node

    batch_size = int(total_batch_size / ngpus_per_node)    
    workers = int((total_workers + ngpus_per_node - 1) / ngpus_per_node)

    model = NeuralNetwork()

    device = torch.device("npu")

    train_sampler = torch.utils.data.distributed.DistributedSampler(training_data)
    test_sampler = torch.utils.data.distributed.DistributedSampler(test_data)

    train_loader = torch.utils.data.DataLoader(
        training_data, batch_size=batch_size, shuffle=(train_sampler is None),
        num_workers=workers, pin_memory=False, sampler=train_sampler, drop_last=True)

    val_loader = torch.utils.data.DataLoader(
        test_data, batch_size=batch_size, shuffle=(test_sampler is None),
        num_workers=workers, pin_memory=False, sampler=test_sampler, drop_last=True)

    loc = 'npu:{}'.format(gpu)
    model = model.to(loc)
    criterion = nn.CrossEntropyLoss().to(loc)
    optimizer = torch.optim.SGD(model.parameters(), lr=1e-3)

    model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[gpu])

    for epoch in range(start_epoch, end_epoch):
        print("curr epoch: ", epoch)
        train_sampler.set_epoch(epoch)
        train(train_loader, model, criterion, optimizer, epoch, gpu)
......    # train函数代码同shell脚本方式

Python方式

def main_worker(gpu, ngpus_per_node, args):

    start_epoch = 0
    end_epoch = 5
    args.gpu = args.local_rank    # 任务拉起后，local_rank自动获得device号

    ddp_setup(args.gpu, args.world_size)
......    # 其余代码同shell脚本方式

配置传参逻辑

在模型脚本中，根据拉起方式不同，需要传入不同的参数，传参配置逻辑如下（此处使用argparse逻辑）：

shell脚本/torchrun/torch_npu_run方式

if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(description='simple distributed training job')
    parser.add_argument('--batch_size', default=512, type=int, help='Input batch size on each device (default: 32)')
    parser.add_argument('--gpu', default=None, type=int,
                    help='GPU id to use.')
    args = parser.parse_args()
    world_size = torch.npu.device_count()
    args.world_size = world_size
    main(args.world_size, args.batch_size)

mp.spawn方式

if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(description='simple distributed training job')
    parser.add_argument('--batch_size', default=512, type=int, help='Input batch size on each device (default: 32)')
    args = parser.parse_args()
    world_size = torch.npu.device_count()
    args.world_size = world_size
    main(args.world_size, args.batch_size)

Python方式

if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(description='simple distributed training job')
    parser.add_argument('--batch_size', default=512, type=int, help='Input batch size on each device (default: 32)')
    parser.add_argument('--gpu', default=None, type=int,
                    help='GPU id to use.')
    parser.add_argument("--local_rank", default=-1, type=int)   # local_rank用于自动获取device号。使用mp.spawn方式与shell方式启动时需删除此项
    args = parser.parse_args()
    world_size = torch.npu.device_count()
    args.world_size = world_size
    main(args.world_size, args.batch_size, args)    # 需将Python拉起命令中设置的参数传入main函数

拉起训练

我们给出了每种方式的拉起命令示例，用户可根据实际情况自行更改。

shell脚本方式

export HCCL_WHITELIST_DISABLE=1
RANK_ID_START=0
WORLD_SIZE=8
for((RANK_ID=$RANK_ID_START;RANK_ID<$((WORLD_SIZE+RANK_ID_START));RANK_ID++));
do
    echo "Device ID: $RANK_ID"
    export LOCAL_RANK=$RANK_ID
    python3 ddp_test_shell.py &
done
wait

mp.spawn方式

export HCCL_WHITELIST_DISABLE=1
python3 ddp_test_spwn.py

Python方式

# master_addr和master_port参数需用户根据实际情况设置
export HCCL_WHITELIST_DISABLE=1
python3 -m torch.distributed.launch --nproc_per_node 8 --master_addr localhost  --master_port *** ddp_test.py

torchrun方式

export HCCL_WHITELIST_DISABLE=1
torchrun --standalone --nnodes=1 --nproc_per_node=8 ddp_test_shell.py

torch_npu_run方式

export HCCL_WHITELIST_DISABLE=1
export MASTER_IP_ADDR=** # 将**填写node_rank0的IP地址
export MASTER_PORT=** # 将**填写为一个空闲的tcp端口号
torch_npu_run --rdzv_backend=parallel --master_addr=$MASTER_IP_ADDR \
--master_port=$MASTER_PORT --nnodes=8 --nproc_per_node=8 ddp_test_shell.py

当屏幕打印/定向日志中出现模型加载、训练等正常运行日志时，说明拉起多卡训练成功。

父主题： 附录