Les traductions sont fournies par des outils de traduction automatique. En cas de conflit entre le contenu d'une traduction et celui de la version originale en anglais, la version anglaise prévaudra.

Exécuter une tâche d'entraînement parallèle sur un modèle SageMaker distribué avec le parallélisme tensoriel

Dans cette section, vous allez apprendre :

Pour en savoir plus sur les smdistributed.modelparallel modules, consultez les API parallèles aux SageMaker modèles dans la documentation du SDK SageMaker Python.

Parallélisme de tenseur seul

Voici un exemple d'option d'entraînement distribué permettant d'activer uniquement le parallélisme de tenseur, sans parallélisme de pipeline. Configurez les smp_options dictionnaires mpi_options et pour spécifier les options de formation distribuées au SageMaker PyTorch ModelTrainer.

Configurez un SageMaker PyTorch ModelTrainer

from sagemaker.train.configs import SourceCode, Compute, InputData
from sagemaker.core import image_uris

mpi_options = {
    "enabled" : True,
    "processes_per_host" : 8,               # 8 processes
    "custom_mpi_options" : "--mca btl_vader_single_copy_mechanism none "
}
               
smp_options = {
    "enabled":True,
    "parameters": {
        "pipeline_parallel_degree": 1,    # alias for "partitions"
        "placement_strategy": "cluster",
        "tensor_parallel_degree": 4,      # tp over 4 devices
        "ddp": True
    }
}
              
# Retrieve the training image for the desired PyTorch version
training_image = image_uris.retrieve(
    framework="pytorch",
    region="us-west-2",
    version='1.13.1',
    py_version='py36',
    instance_type='ml.p3.16xlarge',
    image_scope="training"
)

smp_model_trainer = ModelTrainer(
    training_image=training_image,
    source_code=SourceCode(entry_script='your_training_script.py'),
    role=role,
    compute=Compute(
        instance_type='ml.p3.16xlarge',
        instance_count=1
    ),
    sagemaker_session=sagemaker_session,
    distribution={
        "smdistributed": {"modelparallel": smp_options},
        "mpi": mpi_options
    },
    base_job_name="SMD-MP-demo",
)

smp_model_trainer.train(input_data_config=[
    InputData(channel_name="training", data_source='s3://my_bucket/my_training_data/')
])

Adaptez votre script PyTorch d'entraînement

L'exemple de script d'entraînement suivant montre comment adapter la bibliothèque de parallélisme du SageMaker modèle à un script d'entraînement. Dans cet exemple, on suppose que le script est nommé your_training_script.py.

import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchnet.dataset import SplitDataset
from torchvision import datasets

import smdistributed.modelparallel.torch as smp

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.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 = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.fc2(x)
        return F.log_softmax(x, 1)

def train(model, device, train_loader, optimizer):
    model.train()
    for batch_idx, (data, target) in enumerate(train_loader):
        # smdistributed: Move input tensors to the GPU ID used by
        # the current process, based on the set_device call.
        data, target = data.to(device), target.to(device)
        optimizer.zero_grad()
        output = model(data)
        loss = F.nll_loss(output, target, reduction="mean")
        loss.backward()
        optimizer.step()

# smdistributed: Initialize the backend
smp.init()

# smdistributed: Set the device to the GPU ID used by the current process.
# Input tensors should be transferred to this device.
torch.cuda.set_device(smp.local_rank())
device = torch.device("cuda")

# smdistributed: Download only on a single process per instance.
# When this is not present, the file is corrupted by multiple processes trying
# to download and extract at the same time
if smp.local_rank() == 0:
    dataset = datasets.MNIST("../data", train=True, download=False)
smp.barrier()

# smdistributed: Shard the dataset based on data parallel ranks
if smp.dp_size() > 1:
    partitions_dict = {f"{i}": 1 / smp.dp_size() for i in range(smp.dp_size())}
    dataset = SplitDataset(dataset, partitions=partitions_dict)
    dataset.select(f"{smp.dp_rank()}")

train_loader = torch.utils.data.DataLoader(dataset, batch_size=64)

# smdistributed: Enable tensor parallelism for all supported modules in the model
# i.e., nn.Linear in this case. Alternatively, we can use
# smp.set_tensor_parallelism(model.fc1, True)
# to enable it only for model.fc1
with smp.tensor_parallelism():
    model = Net()

# smdistributed: Use the DistributedModel wrapper to distribute the
# modules for which tensor parallelism is enabled
model = smp.DistributedModel(model)

optimizer = optim.AdaDelta(model.parameters(), lr=4.0)
optimizer = smp.DistributedOptimizer(optimizer)

train(model, device, train_loader, optimizer)

Parallélisme de tenseur associé au parallélisme de pipeline

Voici un exemple d'option d'entraînement distribuée qui permet le parallélisme des tenseurs combiné au parallélisme des pipelines. Configurez les smp_options paramètres mpi_options et pour spécifier les options de parallèle du modèle avec le parallélisme des tenseurs lorsque vous configurez un. SageMaker PyTorch ModelTrainer

Configurez un SageMaker PyTorch ModelTrainer

mpi_options = {
    "enabled" : True,
    "processes_per_host" : 8,               # 8 processes
    "custom_mpi_options" : "--mca btl_vader_single_copy_mechanism none "
}
               
smp_options = {
    "enabled":True,
    "parameters": {
    "microbatches": 4,
        "pipeline_parallel_degree": 2,    # alias for "partitions"
        "placement_strategy": "cluster",
        "tensor_parallel_degree": 2,      # tp over 2 devices
        "ddp": True
    }
}
              
# Retrieve the training image for the desired PyTorch version
training_image = image_uris.retrieve(
    framework="pytorch",
    region="us-west-2",
    version='1.13.1',
    py_version='py36',
    instance_type='ml.p3.16xlarge',
    image_scope="training"
)

smp_model_trainer = ModelTrainer(
    training_image=training_image,
    source_code=SourceCode(entry_script='your_training_script.py'),
    role=role,
    compute=Compute(
        instance_type='ml.p3.16xlarge',
        instance_count=1
    ),
    sagemaker_session=sagemaker_session,
    distribution={
        "smdistributed": {"modelparallel": smp_options},
        "mpi": mpi_options
    },
    base_job_name="SMD-MP-demo",
)

smp_model_trainer.train(input_data_config=[
    InputData(channel_name="training", data_source='s3://my_bucket/my_training_data/')
])

Adaptez votre script PyTorch d'entraînement

L'exemple de script d'entraînement suivant montre comment adapter la bibliothèque de parallélisme du SageMaker modèle à un script d'entraînement. Notez que le script d'entraînement inclut désormais le décorateur smp.step :

import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchnet.dataset import SplitDataset
from torchvision import datasets

import smdistributed.modelparallel.torch as smp

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.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 = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.fc2(x)
        return F.log_softmax(x, 1)


# smdistributed: Define smp.step. Return any tensors needed outside.
@smp.step
def train_step(model, data, target):
    output = model(data)
    loss = F.nll_loss(output, target, reduction="mean")
    model.backward(loss)
    return output, loss

def train(model, device, train_loader, optimizer):
    model.train()
    for batch_idx, (data, target) in enumerate(train_loader):
        # smdistributed: Move input tensors to the GPU ID used by
        # the current process, based on the set_device call.
        data, target = data.to(device), target.to(device)
        optimizer.zero_grad()
        # Return value, loss_mb is a StepOutput object
        _, loss_mb = train_step(model, data, target)

        # smdistributed: Average the loss across microbatches.
        loss = loss_mb.reduce_mean()

        optimizer.step()

# smdistributed: Initialize the backend
smp.init()

# smdistributed: Set the device to the GPU ID used by the current process.
# Input tensors should be transferred to this device.
torch.cuda.set_device(smp.local_rank())
device = torch.device("cuda")

# smdistributed: Download only on a single process per instance.
# When this is not present, the file is corrupted by multiple processes trying
# to download and extract at the same time
if smp.local_rank() == 0:
    dataset = datasets.MNIST("../data", train=True, download=False)
smp.barrier()

# smdistributed: Shard the dataset based on data parallel ranks
if smp.dp_size() > 1:
    partitions_dict = {f"{i}": 1 / smp.dp_size() for i in range(smp.dp_size())}
    dataset = SplitDataset(dataset, partitions=partitions_dict)
    dataset.select(f"{smp.dp_rank()}")

# smdistributed: Set drop_last=True to ensure that batch size is always divisible
# by the number of microbatches
train_loader = torch.utils.data.DataLoader(dataset, batch_size=64, drop_last=True)

model = Net()

# smdistributed: enable tensor parallelism only for model.fc1
smp.set_tensor_parallelism(model.fc1, True)

# smdistributed: Use the DistributedModel container to provide the model
# to be partitioned across different ranks. For the rest of the script,
# the returned DistributedModel object should be used in place of
# the model provided for DistributedModel class instantiation.
model = smp.DistributedModel(model)

optimizer = optim.AdaDelta(model.parameters(), lr=4.0)
optimizer = smp.DistributedOptimizer(optimizer)

train(model, device, train_loader, optimizer)