Run training on Amazon SageMaker

This guide will show you how to train a πŸ€— Transformers model with the HuggingFace SageMaker Python SDK. Learn how to:

Installation and setup

Before you can train a πŸ€— Transformers model with SageMaker, you need to sign up for an AWS account. If you don't have an AWS account yet, learn more here.

Once you have an AWS account, get started using one of the following:

To start training locally, you need to setup an appropriate IAM role.

Upgrade to the latest sagemaker version:

pip install sagemaker --upgrade

SageMaker environment

Setup your SageMaker environment as shown below:

import sagemaker
sess = sagemaker.Session()
role = sagemaker.get_execution_role()

Note: The execution role is only available when running a notebook within SageMaker. If you run get_execution_role in a notebook not on SageMaker, expect a region error.

Local environment

Setup your local environment as shown below:

import sagemaker
import boto3

iam_client = boto3.client('iam')
role = iam_client.get_role(RoleName='role-name-of-your-iam-role-with-right-permissions')['Role']['Arn']
sess = sagemaker.Session()

Prepare a πŸ€— Transformers fine-tuning script

Our training script is very similar to a training script you might run outside of SageMaker. However, you can access useful properties about the training environment through various environment variables (see here for a complete list), such as:

  • SM_MODEL_DIR: A string representing the path to which the training job writes the model artifacts. After training, artifacts in this directory are uploaded to S3 for model hosting. SM_MODEL_DIR is always set to /opt/ml/model.

  • SM_NUM_GPUS: An integer representing the number of GPUs available to the host.

  • SM_CHANNEL_XXXX: A string representing the path to the directory that contains the input data for the specified channel. For example, when you specify train and test in the Hugging Face Estimator fit method, the environment variables are set to SM_CHANNEL_TRAIN and SM_CHANNEL_TEST.

The hyperparameters defined in the Hugging Face Estimator are passed as named arguments and processed by ArgumentParser().

import transformers
import datasets
import argparse
import os

if __name__ == "__main__":

    parser = argparse.ArgumentParser()

    # hyperparameters sent by the client are passed as command-line arguments to the script
    parser.add_argument("--epochs", type=int, default=3)
    parser.add_argument("--per_device_train_batch_size", type=int, default=32)
    parser.add_argument("--model_name_or_path", type=str)

    # data, model, and output directories
    parser.add_argument("--model-dir", type=str, default=os.environ["SM_MODEL_DIR"])
    parser.add_argument("--training_dir", type=str, default=os.environ["SM_CHANNEL_TRAIN"])
    parser.add_argument("--test_dir", type=str, default=os.environ["SM_CHANNEL_TEST"])

Note that SageMaker doesn’t support argparse actions. For example, if you want to use a boolean hyperparameter, specify type as bool in your script and provide an explicit True or False value.

Look here for a complete example of a πŸ€— Transformers training script.

Training Output Management

If output_dir in the TrainingArguments is set to '/opt/ml/model' the Trainer saves all training artifacts, including logs, checkpoints, and models. Amazon SageMaker archives the whole '/opt/ml/model' directory as model.tar.gz and uploads it at the end of the training job to Amazon S3. Depending on your Hyperparameters and TrainingArguments this could lead to a large artifact (> 5GB), which can slow down deployment for Amazon SageMaker Inference. You can control how checkpoints, logs, and artifacts are saved by customization the TrainingArguments. For example by providing save_total_limit as TrainingArgument you can control the limit of the total amount of checkpoints. Deletes the older checkpoints in output_dir if new ones are saved and the maximum limit is reached.

In addition to the options already mentioned above, there is another option to save the training artifacts during the training session. Amazon SageMaker supports Checkpointing, which allows you to continuously save your artifacts during training to Amazon S3 rather than at the end of your training. To enable Checkpointing you need to provide the checkpoint_s3_uri parameter pointing to an Amazon S3 location in the HuggingFace estimator and set output_dir to /opt/ml/checkpoints. Note: If you set output_dir to /opt/ml/checkpoints make sure to call trainer.save_model("/opt/ml/model") or model.save_pretrained("/opt/ml/model")/tokenizer.save_pretrained("/opt/ml/model") at the end of your training to be able to deploy your model seamlessly to Amazon SageMaker for Inference.

Create a Hugging Face Estimator

Run πŸ€— Transformers training scripts on SageMaker by creating a Hugging Face Estimator. The Estimator handles end-to-end SageMaker training. There are several parameters you should define in the Estimator:

  1. entry_point specifies which fine-tuning script to use.
  2. instance_type specifies an Amazon instance to launch. Refer here for a complete list of instance types.
  3. hyperparameters specifies training hyperparameters. View additional available hyperparameters here.

The following code sample shows how to train with a custom script train.py with three hyperparameters (epochs, per_device_train_batch_size, and model_name_or_path):

from sagemaker.huggingface import HuggingFace


# hyperparameters which are passed to the training job
hyperparameters={'epochs': 1,
                 'per_device_train_batch_size': 32,
                 'model_name_or_path': 'distilbert-base-uncased'
                 }

# create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='train.py',
        source_dir='./scripts',
        instance_type='ml.p3.2xlarge',
        instance_count=1,
        role=role,
        transformers_version='4.4',
        pytorch_version='1.6',
        py_version='py36',
        hyperparameters = hyperparameters
)

If you are running a TrainingJob locally, define instance_type='local' or instance_type='local-gpu' for GPU usage. Note that this will not work with SageMaker Studio.

Execute training

Start your TrainingJob by calling fit on a Hugging Face Estimator. Specify your input training data in fit. The input training data can be a:

  • S3 URI such as s3://my-bucket/my-training-data.
  • FileSystemInput for Amazon Elastic File System or FSx for Lustre. See here for more details about using these file systems as input.

Call fit to begin training:

huggingface_estimator.fit(
  {'train': 's3://sagemaker-us-east-1-558105141721/samples/datasets/imdb/train',
   'test': 's3://sagemaker-us-east-1-558105141721/samples/datasets/imdb/test'}
)

SageMaker starts and manages all the required EC2 instances and initiates the TrainingJob by running:

/opt/conda/bin/python train.py --epochs 1 --model_name_or_path distilbert-base-uncased --per_device_train_batch_size 32

Access trained model

Once training is complete, you can access your model through the AWS console or download it directly from S3.

from sagemaker.s3 import S3Downloader

S3Downloader.download(
    s3_uri=huggingface_estimator.model_data, # S3 URI where the trained model is located
    local_path='.',                          # local path where *.targ.gz is saved
    sagemaker_session=sess                   # SageMaker session used for training the model
)

Distributed training

SageMaker provides two strategies for distributed training: data parallelism and model parallelism. Data parallelism splits a training set across several GPUs, while model parallelism splits a model across several GPUs.

Data parallelism

The Hugging Face Trainer supports SageMaker's data parallelism library. If your training script uses the Trainer API, you only need to define the distribution parameter in the Hugging Face Estimator:

# configuration for running training on smdistributed data parallel
distribution = {'smdistributed':{'dataparallel':{ 'enabled': True }}}

# create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='train.py',
        source_dir='./scripts',
        instance_type='ml.p3dn.24xlarge',
        instance_count=2,
        role=role,
        transformers_version='4.4.2',
        pytorch_version='1.6.0',
        py_version='py36',
        hyperparameters = hyperparameters
        distribution = distribution
)

πŸ““ Open the notebook for an example of how to run the data parallelism library with TensorFlow.

Model parallelism

The Hugging Face [Trainer] also supports SageMaker's model parallelism library. If your training script uses the Trainer API, you only need to define the distribution parameter in the Hugging Face Estimator (see here for more detailed information about using model parallelism):

# configuration for running training on smdistributed model parallel
mpi_options = {
    "enabled" : True,
    "processes_per_host" : 8
}

smp_options = {
    "enabled":True,
    "parameters": {
        "microbatches": 4,
        "placement_strategy": "spread",
        "pipeline": "interleaved",
        "optimize": "speed",
        "partitions": 4,
        "ddp": True,
    }
}

distribution={
    "smdistributed": {"modelparallel": smp_options},
    "mpi": mpi_options
}

 # create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='train.py',
        source_dir='./scripts',
        instance_type='ml.p3dn.24xlarge',
        instance_count=2,
        role=role,
        transformers_version='4.4.2',
        pytorch_version='1.6.0',
        py_version='py36',
        hyperparameters = hyperparameters,
        distribution = distribution
)

πŸ““ Open the notebook for an example of how to run the model parallelism library.

Spot instances

The Hugging Face extension for the SageMaker Python SDK means we can benefit from fully-managed EC2 spot instances. This can help you save up to 90% of training costs!

Note: Unless your training job completes quickly, we recommend you use checkpointing with managed spot training. In this case, you need to define the checkpoint_s3_uri.

Set use_spot_instances=True and define your max_wait and max_run time in the Estimator to use spot instances:

# hyperparameters which are passed to the training job
hyperparameters={'epochs': 1,
                 'train_batch_size': 32,
                 'model_name':'distilbert-base-uncased',
                 'output_dir':'/opt/ml/checkpoints'
                 }

# create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='train.py',
        source_dir='./scripts',
        instance_type='ml.p3.2xlarge',
        instance_count=1,
        checkpoint_s3_uri=f's3://{sess.default_bucket()}/checkpoints'
        use_spot_instances=True,
        # max_wait should be equal to or greater than max_run in seconds
        max_wait=3600,
        max_run=1000,
        role=role,
        transformers_version='4.4',
        pytorch_version='1.6',
        py_version='py36',
        hyperparameters = hyperparameters
)

# Training seconds: 874
# Billable seconds: 262
# Managed Spot Training savings: 70.0%

πŸ““ Open the notebook for an example of how to use spot instances.

Git repository

The Hugging Face Estimator can load a training script stored in a GitHub repository. Provide the relative path to the training script in entry_point and the relative path to the directory in source_dir.

If you are using git_config to run the πŸ€— Transformers example scripts, you need to configure the correct 'branch' in transformers_version (e.g. if you use transformers_version='4.4.2 you have to use 'branch':'v4.4.2').

Tip: Save your model to S3 by setting output_dir=/opt/ml/model in the hyperparameter of your training script.

# configure git settings
git_config = {'repo': 'https://github.com/huggingface/transformers.git','branch': 'v4.4.2'} # v4.4.2 refers to the transformers_version you use in the estimator

 # create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='run_glue.py',
        source_dir='./examples/pytorch/text-classification',
        git_config=git_config,
        instance_type='ml.p3.2xlarge',
        instance_count=1,
        role=role,
        transformers_version='4.4',
        pytorch_version='1.6',
        py_version='py36',
        hyperparameters=hyperparameters
)

SageMaker metrics

SageMaker metrics automatically parses training job logs for metrics and sends them to CloudWatch. If you want SageMaker to parse the logs, you must specify the metric's name and a regular expression for SageMaker to use to find the metric.

# define metrics definitions
metric_definitions = [
    {"Name": "train_runtime", "Regex": "train_runtime.*=\D*(.*?)$"},
    {"Name": "eval_accuracy", "Regex": "eval_accuracy.*=\D*(.*?)$"},
    {"Name": "eval_loss", "Regex": "eval_loss.*=\D*(.*?)$"},
]

# create the Estimator
huggingface_estimator = HuggingFace(
        entry_point='train.py',
        source_dir='./scripts',
        instance_type='ml.p3.2xlarge',
        instance_count=1,
        role=role,
        transformers_version='4.4',
        pytorch_version='1.6',
        py_version='py36',
        metric_definitions=metric_definitions,
        hyperparameters = hyperparameters)

πŸ““ Open the notebook for an example of how to capture metrics in SageMaker.