一、Code Sample
from __future__ import print_function import os import time import argparse import torch import torch.nn as nn import torch.nn.functional as F import torch.optim as optim from torchvision import datasets, transforms from torch.optim.lr_scheduler import StepLR import torch.distributed as dist def init_distributed_mode(args): '''initilize DDP ''' if "RANK" in os.environ and "WORLD_SIZE" in os.environ: args.rank = int(os.environ["RANK"]) args.world_size = int(os.environ["WORLD_SIZE"]) args.gpu = int(os.environ["LOCAL_RANK"]) elif "SLURM_PROCID" in os.environ: args.rank = int(os.environ["SLURM_PROCID"]) args.gpu = args.rank % torch.cuda.device_count() elif hasattr(args, "rank"): pass else: print("Not using distributed mode") args.distributed = False return args.distributed = True torch.cuda.set_device(args.gpu) args.dist_backend = "nccl" # args.dist_backend = "gloo" print(f"| distributed init (rank {args.rank}): {args.dist_url}, local rank:{args.gpu}, world size:{args.world_size}", flush=True) dist.init_process_group( backend=args.dist_backend, init_method=args.dist_url, world_size=args.world_size, rank=args.rank ) class Net(nn.Module): def __init__(self): super(Net, self).__init__() self.conv1 = nn.Conv2d(1, 32, 3, 1) self.conv2 = nn.Conv2d(32, 64, 3, 1) self.dropout1 = nn.Dropout(0.25) self.dropout2 = nn.Dropout(0.5) self.fc1 = nn.Linear(9216, 128) self.fc2 = nn.Linear(128, 10) def forward(self, x): x = self.conv1(x) x = F.relu(x) x = self.conv2(x) x = F.relu(x) x = F.max_pool2d(x, 2) x = self.dropout1(x) x = torch.flatten(x, 1) x = self.fc1(x) x = F.relu(x) x = self.dropout2(x) x = self.fc2(x) output = F.log_softmax(x, dim=1) return output def train(args, model, device, train_loader, optimizer, epoch): model.train() for batch_idx, (data, target) in enumerate(train_loader): data, target = data.to(device), target.to(device) optimizer.zero_grad() output = model(data) loss = F.nll_loss(output, target) loss.backward() optimizer.step() if args.distributed: if dist.get_rank() == 0: if batch_idx % args.log_interval == 0: print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format( epoch, dist.get_world_size() * batch_idx * len(data), len(train_loader.dataset), 100. * batch_idx / len(train_loader), loss.item())) else: if batch_idx % args.log_interval == 0: print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format( epoch, batch_idx * len(data), len(train_loader.dataset), 100. * batch_idx / len(train_loader), loss.item())) if args.dry_run: break def test(model, device, test_loader): model.eval() test_loss = 0 correct = 0 with torch.no_grad(): for data, target in test_loader: data, target = data.to(device), target.to(device) output = model(data) test_loss += F.nll_loss(output, target, reduction='sum').item() # sum up batch loss pred = output.argmax(dim=1, keepdim=True) # get the index of the max log-probability correct += pred.eq(target.view_as(pred)).sum().item() test_loss /= len(test_loader.dataset) print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format( test_loss, correct, len(test_loader.dataset), 100. * correct / len(test_loader.dataset))) def main(): # Training settings parser = argparse.ArgumentParser(description='PyTorch MNIST Example') parser.add_argument('--batch-size', type=int, default=64, metavar='N', help='input batch size for training (default: 64)') parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N', help='input batch size for testing (default: 1000)') parser.add_argument('--epochs', type=int, default=14, metavar='N', help='number of epochs to train (default: 14)') parser.add_argument('--lr', type=float, default=1.0, metavar='LR', help='learning rate (default: 1.0)') parser.add_argument('--gamma', type=float, default=0.7, metavar='M', help='Learning rate step gamma (default: 0.7)') parser.add_argument('--no-cuda', action='store_true', default=False, help='disables CUDA training') parser.add_argument('--dry-run', action='store_true', default=False, help='quickly check a single pass') parser.add_argument('--seed', type=int, default=1, metavar='S', help='random seed (default: 1)') parser.add_argument('--log-interval', type=int, default=10, metavar='N', help='how many batches to wait before logging training status') parser.add_argument('--save-model', action='store_true', default=False, help='For Saving the current Model') parser.add_argument('--local_rank', type=int, help='local rank, will passed by ddp') parser.add_argument("--world-size", default=1, type=int, help="number of distributed processes") parser.add_argument("--dist-url", default="env://", type=str, help="url used to set up distributed training") args = parser.parse_args() use_cuda = not args.no_cuda and torch.cuda.is_available() init_distributed_mode(args) torch.manual_seed(args.seed) device = torch.device("cuda" if use_cuda else "cpu") train_kwargs = {'batch_size': args.batch_size} test_kwargs = {'batch_size': args.test_batch_size} if use_cuda: cuda_kwargs = {'num_workers': 1, 'pin_memory': True, } train_kwargs.update(cuda_kwargs) test_kwargs.update(cuda_kwargs) transform=transforms.Compose([ transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,)) ]) dataset_train = datasets.MNIST('./data', train=True, download=True, transform=transform) dataset_val = datasets.MNIST('./data', train=False, transform=transform) if args.distributed: train_sampler = torch.utils.data.distributed.DistributedSampler(dataset_train, shuffle=True) else: train_sampler = torch.utils.data.RandomSampler(dataset_train) test_sampler = torch.utils.data.SequentialSampler(dataset_val) train_loader = torch.utils.data.DataLoader(dataset_train, sampler = train_sampler, **train_kwargs) test_loader = torch.utils.data.DataLoader(dataset_val, sampler = test_sampler, **test_kwargs) model = Net().to(device) model_without_ddp = model if args.distributed: model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu]) model_without_ddp = model.module optimizer = optim.Adadelta(model.parameters(), lr=args.lr) scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma) for epoch in range(1, args.epochs + 1): if args.distributed: train_sampler.set_epoch(epoch) train(args, model, device, train_loader, optimizer, epoch) if args.distributed: # Only run validation on GPU 0 process, for simplity, so we do not run validation on multi gpu. if dist.get_rank() == 0: test(model_without_ddp, device, test_loader) else: test(model, device, test_loader) scheduler.step() if args.save_model: if args.distributed: if dist.get_rank() == 0: # only save model on GPU0 process. torch.save(model.state_dict(), f"mnist_cnn.pt") else: torch.save(model.state_dict(), f"mnist_cnn_.pt") if __name__ == '__main__': start = time.time() main() print(f'Total cost time:{time.time() - start} ms')
二、将python文件上传的oss
如果有配置数据源nas等，也可以直接将文件上传到nas中
三、创建任务
wget https://********.oss-cn-hangzhou.aliyuncs.com/mnist-ddp.py && python -m torch.distributed.launch --master_addr=$MASTER_ADDR --master_port=$MASTER_PORT --nproc_per_node=1 --nnodes=$WORLD_SIZE --node_rank=$RANK mnist-ddp.py
四、察看运行结果
参考链接
提交分布式DLC任务
