Monthly Archives: June 2023

Amazon SageMaker Automatic Model Tuning has introduced Autotune, a new feature to automatically choose hyperparameters on your behalf. This provides an accelerated and more efficient way to find hyperparameter ranges, and can provide significant optimized budget and time management for your automatic model tuning jobs.

In this post, we discuss this new capability and some of the benefits it brings.

Hyperparameter overview

When training any machine learning (ML) model, you are generally dealing with three types of data: input data (also called the training data), model parameters, and hyperparameters. You use the input data to train your model, which in effect learns your model parameters. During the training process, your ML algorithms are trying to find the optimal model parameters based on data while meeting the goals of your objective function. For example, when a neural network is trained, the weight of the network nodes is learned from the training, and indicates how much impact it has on the final prediction. These weights are the model parameters.

Hyperparameters, on the other hand, are parameters of a learning algorithm and not the model itself. The number of hidden layers and the number of nodes are some of the examples of hyperparameters you can set for a neural network. The difference between model parameters and hyperparameters is that model parameters are learned during the training process, whereas hyperparameters are set prior to the training and remain constant during the training process.

Pain points

SageMaker automatic model tuning, also called hyperparameter tuning, runs many training jobs on your dataset using a range of hyperparameters that you specify. It can accelerate your productivity by trying many variations of a model. It looks for the best model automatically by focusing on the most promising combinations of hyperparameter values within the ranges that you specify. However, to get good results, you must choose the right ranges to explore.

But how do you know what the right range is to begin with? With hyperparameter tuning jobs, we are assuming that the optimal set of hyperparameters lies within the range that we specified. What happens if the chosen range is not right, and the optimal hyperparameter actually falls outside of the range?

Choosing the right hyperparameters requires experience with the ML technique you are using and understanding how its hyperparameters behave. It’s important to understand the hyperparameter implications because every hyperparameter that you choose to tune has the potential to increase the number of trials required for a successful tuning job. You need to strike an optimal trade-off between resources allocated to the tuning job and achieving the goals you’ve set.

The SageMaker Automatic Model Tuning team is constantly innovating on behalf of our customers to optimize their ML workloads. AWS recently announced support of new completion criteria for hyperparameter optimization: the max runtime criteria, which is a budget control completion criteria that can be used to bound cost and runtime. Desired target metrics, improvement monitoring, and convergence detection monitors the performance of the model and assists with early stopping if the models don’t improve after a defined number of training jobs. Autotune is a new feature of automatic model tuning that helps save you time and reduce wasted resources on finding optimal hyperparameter ranges.

Benefits of Autotune and how automatic model tuning alleviates those pain points

Autotune is a new configuration in the CreateHyperParameterTuningJob API and in the HyperparameterTuner SageMaker Python SDK that alleviates the need to specify the hyperparameter ranges, tuning strategy, objective metrics, or the number of jobs that were required as part of the job definition. Autotune automatically chooses the optimal configurations for your tuning job, helps prevent wasted resources, and accelerates productivity.

The following example showcases how many of the parameters are not necessary when using Autotune.

The following code creates a hyperparameter tuner using the SageMaker Python SDK without Autotune:

estimator = PyTorch(
    entry_point="mnist.py",
    instance_type="ml.p4d.24xlarge",
    hyperparameters={
        "epochs": 1, "backend": "gloo"
    },
)

tuner = HyperparameterTuner(
    estimator, 
    objective_metric_name='validation:rmse',
    objective_type='Minimize',
    hyperparameter_ranges = {
        "lr": ContinuousParameter(0.001, 0.1),
        "batch-size": CategoricalParameter([32, 64, 128, 256, 512])
    },
    metric_definitions=[{...}],
    max_jobs=10,
    strategy="Random"
)

tuner.fit(...)

The following example showcases how many of the parameters are not necessary when using Autotune:

estimator = PyTorch(
    entry_point="mnist.py",
    instance_type="ml.p4d.24xlarge",
    hyperparameters={
        "epochs": 1, "backend": "gloo", "lr": 0.01, "batch-size": 32
    },
)
tuner = HyperparameterTuner(
    estimator, 
    objective_metric_name='validation:rmse',
    objective_type='Minimize', 
    autotune=True
)

If you are using API, the equivalent code would be as follows:

create_hyper_parameter_tuning_job(
    HyperParameterTuningJobName=tuning_job_name,
    HyperParameterTuningJobConfig=tuning_job_config,
    TrainingJobDefinition=training_job_definition,
    Autotune={'Mode': 'Enabled'},
)

The code example illustrates some of the key benefits of Autotune:

• A key choice for a tuning job is which hyperparameters to tune and their ranges. Autotune makes this choice for you based on a list of hyperparameters that you provide. Using the previous example, the hyperparameters that Autotune can choose to be tunable are lr and batch-size.
• Autotune will automatically select the hyperparameter ranges on your behalf. Autotune uses best practices as well as internal benchmarks for selecting the appropriate ranges.
• Autotune automatically selects the strategy on how to choose the combinations of hyperparameter values to use for the training job.
• Early stopping is enabled by default when using Autotune. When using early stopping, SageMaker stops training jobs launched by the hyperparameter tuning job when they are unlikely to perform better than previously completed training jobs to avoid additional resource utilization.
• Maximum expected resources to be consumed by the tuning job (parallel jobs, max runtime, and so on) will be calculated and set in the tuning job record as soon as the tuning job is created. Such reserved resources will not increase during the course of the tuning job; this will maintain an upper bound of cost and duration of the tuning job that is easily predictable by the user. A max runtime of 48 hours will be used by default.

You can override any settings chosen automatically by Autotune. As an example, if you specify your own hyperparameter ranges, those will be used alongside the inferred ranges. Any user-specified hyperparameter range will take precedence over the same named inferred ranges:

estimator = PyTorch(
    ...
    hyperparameters={
        "epochs": 100, "backend": "gloo", "lr": 0.01, "beta1": 0.8
    }

tuner = HyperparameterTuner(
    ...
    hyperparameter_ranges = {
        "lr": ContinuousParameter(0.001, 0.01) # takes precedence over inferred "lr"
    }

Autotune generates a set of settings as part of the tuning job. Any customer-specified settings that have the same name will override the Autotune-selected settings. Any customer-provided settings (that aren’t the same as the named Autotune settings) are added in addition to the Autotune-selected settings.

Inspecting parameters chosen by Autotune

Autotune reduces the time you would normally have spent in deciding on the initial set of hyperparameters to tune. But how do you get insights into what hyperparameter values Autotune ended up choosing? You can get information about decisions made for you in the description of the running tuning job (in the response of the DescribeHyperParameterTuningJob operation). After you submit a request to create a tuning job, the request is processed, and all missing fields are set by Autotune. All set fields are reported in the DescribeHyperParameterTuningJob operation.

Alternatively, you can inspect HyperparameterTuner class fields to see the settings chosen by Autotune.

The following is an XGBoost example of how you may use the DescribeHyperParameterTuningJob to inspect the hyperparameters chosen by Autotune.

First, we create a tuning job with automatic model tuning:

hyperparameters = {
    "objective": "reg:squarederror",
    "num_round": "50",
    "verbosity": "2",
    "max_depth": "5",  # overlap with ranges is ok when Autotune is enabled
}
estimator = XGBoost(hyperparameters=hyperparameters, ...)

hp_tuner = HyperparameterTuner(estimator, autotune=True)
hp_tuner.fit(wait=False)

After the tuning job is created successfully, we can discover what settings Autotune chose. For example, we can describe the tuning job by the name given by it from hp_tuner:

import boto3 
sm = boto3.client('sagemaker')

response = sm.describe_hyper_parameter_tuning_job(
   HyperParameterTuningJobName=hp_tuner.latest_tuning_job.name
)

print(response)

Then we can inspect the generated response to review the settings chosen by Autotune on our behalf.

If the current tuning job settings are not satisfactory, you can stop the tuning job:

hp_tuner.stop()

Conclusion

SageMaker Automatic Model Tuning allows you to reduce the time to tune a model by automatically searching for the best hyperparameter configuration within the ranges that you specify. However, choosing the right hyperparameter ranges can be a time-consuming process and can have direct implications on your training cost and duration.

In this post, we discussed how you can now use Autotune, a new feature introduced as part of automatic model tuning, to automatically pick an initial set of hyperparameter ranges on your behalf. This can reduce the time it takes for you to get started with your model tuning process. Additionally, you can evaluate the ranges picked by Autotune and adjust them according to your needs.

We also showed how Autotune can automatically pick the optimal parameter settings on your behalf, such as the number of training jobs, the strategy to choose the hyperparameter combinations, and enabling early stopping by default. This can result in significantly optimized budget and time bounds that are easily predictable.

To learn more, refer to Perform Automatic Model Tuning with SageMaker.

About the Authors

Jas Singh is a Senior Solutions Architect helping public sector customers achieve their business outcomes through architecting and implementing innovative and resilient solutions at scale. Jas has over 20 years of experience in designing and implementing mission-critical applications and holds a master’s degree in computer science from Baylor University.

Gopi Mudiyala is a Senior Technical Account Manager at AWS. He helps customers in the Financial Services industry with their operations in AWS. As a machine learning enthusiast, Gopi works to help customers succeed in their ML journey. In his spare time, he likes to play badminton, spend time with family, and travel.

Raviteja Yelamanchili is an Enterprise Solutions Architect with Amazon Web Services based in New York. He works with large financial services enterprise customers to design and deploy highly secure, scalable, reliable, and cost-effective applications on the cloud. He brings over 11 years of risk management, technology consulting, data analytics, and machine learning experience. When he is not helping customers, he enjoys traveling and playing PS5.

Iaroslav Shcherbatyi is a Machine Learning Engineer at AWS. He works mainly on improvements to the Amazon SageMaker platform and helping customers best use its features. In his spare time, he likes to go to gym, do outdoor sports such as ice skating or hiking, and to catch up on new AI research.

Read More

This post is co-written with Philipp Schmid from Hugging Face.

We have all heard about the progress being made in the field of large language models (LLMs) and the ever-growing number of problem sets where LLMs are providing valuable insights. Large models, when trained over massive datasets and several tasks, are also able to generalize well over tasks that they aren’t trained specifically for. Such models are called foundation models, a term first popularized by the Stanford Institute for Human-Centered Artificial Intelligence. Even though these foundation models are able to generalize well, especially with the help of prompt engineering techniques, often the use case is so domain specific, or the task is so different, that the model needs further customization. One approach to improve performance of a large model for a specific domain or task is to further train the model with a smaller, task-specific dataset. Although this approach, known as fine-tuning, successfully improves the accuracy of LLMs, it requires modifying all of the model weights. Fine-tuning is much faster than the pre-training of a model thanks to the much smaller dataset size, but still requires significant computing power and memory. Fine-tuning modifies all the parameter weights of the original model, which makes it expensive and results in a model that is the same size as the original.

To address these challenges, Hugging Face introduced the Parameter-Efficient Fine-Tuning library (PEFT). This library allows you to freeze most of the original model weights and replace or extend model layers by training an additional, much smaller, set of parameters. This makes training much less expensive in terms of required compute and memory.

In this post, we show you how to train the 7-billion-parameter BloomZ model using just a single graphics processing unit (GPU) on Amazon SageMaker, Amazon’s machine learning (ML) platform for preparing, building, training, and deploying high-quality ML models. BloomZ is a general-purpose natural language processing (NLP) model. We use PEFT to optimize this model for the specific task of summarizing messenger-like conversations. The single-GPU instance that we use is a low-cost example of the many instance types AWS provides. Training this model on a single GPU highlights AWS’s commitment to being the most cost-effective provider of AI/ML services.

The code for this walkthrough can be found on the Hugging Face notebooks GitHub repository under the sagemaker/24_train_bloom_peft_lora folder.

Prerequisites

In order to follow along, you should have the following prerequisites:

• An AWS account.
• A Jupyter notebook within Amazon SageMaker Studio or SageMaker notebook instances.
• You will need access to the SageMaker ml.g5.2xlarge instance type, containing a single NVIDIA A10G GPU. On the AWS Management Console, navigate to Service Quotas for SageMaker and request a 1-instance increase for the following quotas: ml.g5.2xlarge for training job usage and ml.g5.2xlarge for training job usage.
• After your requested quotas are applied to your account, you can use the default Studio Python 3 (Data Science) image with a ml.t3.medium instance to run the notebook code snippets. For the full list of available kernels, refer to Available Amazon SageMaker Kernels.

Set up a SageMaker session

Use the following code to set up your SageMaker session:

import sagemaker
import boto3
sess = sagemaker.Session()
# sagemaker session bucket -> used for uploading data, models and logs
# sagemaker will automatically create this bucket if it does not exist
sagemaker_session_bucket=None
if sagemaker_session_bucket is None and sess is not None:
    # set to default bucket if a bucket name is not given
    sagemaker_session_bucket = sess.default_bucket()

try:
    role = sagemaker.get_execution_role()
except ValueError:
    iam = boto3.client('iam')
    role = iam.get_role(RoleName='sagemaker_execution_role')['Role']['Arn']

sess = sagemaker.Session(default_bucket=sagemaker_session_bucket)

print(f"sagemaker role arn: {role}")
print(f"sagemaker bucket: {sess.default_bucket()}")
print(f"sagemaker session region: {sess.boto_region_name}")

Load and prepare the dataset

We use the samsum dataset, a collection of 16,000 messenger-like conversations with summaries. The conversations were created and written down by linguists fluent in English. The following is an example of the dataset:

{
  "id": "13818513",
  "summary": "Amanda baked cookies and will bring Jerry some tomorrow.",
  "dialogue": "Amanda: I baked cookies. Do you want some?rnJerry: Sure!rnAmanda: I'll bring you tomorrow :-)"
}

To train the model, you need to convert the inputs (text) to token IDs. This is done by a Hugging Face Transformers tokenizer. For more information, refer to Chapter 6 of the Hugging Face NLP Course.

Convert the inputs with the following code:

from transformers import AutoTokenizer

model_id="bigscience/bloomz-7b1"

# Load tokenizer of BLOOMZ
tokenized = AutoTokenizer.from_pretrained(model_id)
tokenizer.model_max_length = 2048 # overwrite wrong value

Before starting training, you need to process the data. Once it’s trained, the model will take a set of text messages as the input and generate a summary as the output. You need to format the data as a prompt (the messages) with a correct response (the summary). You also need to chunk examples into longer input sequences to optimize the model training. See the following code:

from random import randint
from itertools import chain
from functools import partial

# custom instruct prompt start
prompt_template = f"Summarize the chat dialogue:n{{dialogue}}n---nSummary:n{{summary}}{{eos_token}}"

# template dataset to add prompt to each sample
def template_dataset(sample):
    sample["text"] = prompt_template.format(dialogue=sample["dialogue"],
                                            summary=sample["summary"],
                                            eos_token=tokenizer.eos_token)
    return sample


# apply prompt template per sample
dataset = dataset.map(template_dataset, remove_columns=list(dataset.features))

print(dataset[randint(0, len(dataset))]["text"])

# empty list to save remainder from batches to use in next batch
remainder = {"input_ids": [], "attention_mask": []}


def chunk(sample, chunk_length=2048):
    # define global remainder variable to save remainder from batches to use in next batch
    global remainder
    # Concatenate all texts and add remainder from previous batch
    concatenated_examples = {k: list(chain(*sample[k])) for k in sample.keys()}
    concatenated_examples = {k: remainder[k] + concatenated_examples[k] for k in concatenated_examples.keys()}
    # get total number of tokens for batch
    batch_total_length = len(concatenated_examples[list(sample.keys())[0]])

    # get max number of chunks for batch
    if batch_total_length >= chunk_length:
        batch_chunk_length = (batch_total_length // chunk_length) * chunk_length

    # Split by chunks of max_len.
    result = {
        k: [t[i : i + chunk_length] for i in range(0, batch_chunk_length, chunk_length)]
        for k, t in concatenated_examples.items()
    }
    # add remainder to global variable for next batch
    remainder = {k: concatenated_examples[k][batch_chunk_length:] for k in concatenated_examples.keys()}
    # prepare labels
    result["labels"] = result["input_ids"].copy()
    return result


# tokenize and chunk dataset
lm_dataset = dataset.map(
    lambda sample: tokenizer(sample["text"]), batched=True, remove_columns=list(dataset.features)
).map(
    partial(chunk, chunk_length=2048),
    batched=True,
)

# Print total number of samples
print(f"Total number of samples: {len(lm_dataset)}")

Now you can use the FileSystem integration to upload the dataset to Amazon Simple Storage Service (Amazon S3):

# save train_dataset to s3
training_input_path = f's3://{sess.default_bucket()}/processed/samsum-sagemaker/train'
lm_dataset.save_to_disk(training_input_path)

print("uploaded data to:")
print(f"training dataset to: {training_input_path}")

In [ ]:
training_input_path="s3://sagemaker-us-east-1-558105141721/processed/samsum-sagemaker/train"

Fine-tune BLOOMZ-7B with LoRA and bitsandbytes int-8 on SageMaker

The Hugging Face BLOOMZ-7B model card indicates its initial training was distributed over 8 nodes with 8 A100 80 GB GPUs and 512 GB memory CPUs each. This computing configuration is not readily accessible, is cost-prohibitive to consumers, and requires expertise in distributed training performance optimization. SageMaker lowers the barriers to replication of this setup through its distributed training libraries; however, the cost of comparable eight on-demand ml.p4de.24xlarge instances would be $376.88 per hour. Furthermore, the fully trained model consumes about 40 GB of memory, which exceeds the available memory of many individual consumer available GPUs and requires strategies to address for large model inferencing. As a result, full fine-tuning of the model for your task over multiple model runs and deployment would require significant compute, memory, and storage costs on hardware that isn’t readily accessible to consumers.

Our goal is to find a way to adapt BLOOMZ-7B to our chat summarization use case in a more accessible and cost-effective way while maintaining accuracy. To enable our model to be fine-tuned on a SageMaker ml.g5.2xlarge instance with a single consumer-grade NVIDIA A10G GPU, we employ two techniques to reduce the compute and memory requirements for fine-tuning: LoRA and quantization.

LoRA (Low Rank Adaptation) is a technique that significantly reduces the number of model parameters and associated compute needed for fine-tuning to a new task without a loss in predictive performance. First, it freezes your original model weights and instead optimizes smaller rank-decomposition weight matrices to your new task rather than updating the full weights, and then injects these adapted weights back into the original model. Consequently, fewer weight gradient updates means less compute and GPU memory during fine-tuning. The intuition behind this approach is that LoRA allows LLMs to focus on the most important input and output tokens while ignoring redundant and less important tokens. To deepen your understanding of the LoRA technique, refer to the original paper LoRA: Low-Rank Adaptation of Large Language Models.

In addition to the LoRA technique, you use the bitsanbytes Hugging Face integration LLM.int8() method to quantize out the frozen BloomZ model, or reduce the precision of the weight and bias values, by rounding them from float16 to int8. Quantization reduces the needed memory for BloomZ by about four times, which enables you to fit the model on the A10G GPU instance without a significant loss in predictive performance. To deepen your understanding of how int8 quantization works, its implementation in the bitsandbytes library, and its integration with the Hugging Face Transformers library, see A Gentle Introduction to 8-bit Matrix Multiplication for transformers at scale using Hugging Face Transformers, Accelerate and bitsandbytes.

Hugging Face has made LoRA and quantization accessible across a broad range of transformer models through the PEFT library and its integration with the bitsandbytes library. The create_peft_config() function in the prepared script run_clm.py illustrates their usage in preparing your model for training:

def create_peft_config(model):
    from peft import (
        get_peft_model,
        LoraConfig,
        TaskType,
        prepare_model_for_int8_training,
    )

    peft_config = LoraConfig(
        task_type=TaskType.CAUSAL_LM,
        inference_mode=False,
        r=8, # Lora attention dimension.
        lora_alpha=32, # the alpha parameter for Lora scaling.
        lora_dropout=0.05, # the dropout probability for Lora layers.
        target_modules=["query_key_value"],
    )

    # prepare int-8 model for training
    model = prepare_model_for_int8_training(model)
    model = get_peft_model(model, peft_config)
    model.print_trainable_parameters()
    return model

With LoRA, the output from print_trainable_parameters()indicates we were able to reduce the number of model parameters from 7 billion to 3.9 million. This means that only 5.6% of the original model parameters need to be updated. This significant reduction in compute and memory requirements allows us to fit and train our model on the GPU without issues.

To create a SageMaker training job, you will need a Hugging Face estimator. The estimator handles end-to-end SageMaker training and deployment tasks. SageMaker takes care of starting and managing all the required Amazon Elastic Compute Cloud (Amazon EC2) instances for you. Additionally, it provides the correct Hugging Face training container, uploads the provided scripts, and downloads the data from our S3 bucket into the container at the path /opt/ml/input/data. Then, it starts the training job. See the following code:

import time
# define Training Job Name 
job_name = f'huggingface-peft-{time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime())}'

from sagemaker.huggingface import HuggingFace

# hyperparameters, which are passed into the training job
hyperparameters ={
  'model_id': model_id,                                # pre-trained model
  'dataset_path': '/opt/ml/input/data/training', # path where sagemaker will save training dataset
  'epochs': 3,                                         # number of training epochs
  'per_device_train_batch_size': 1,                    # batch size for training
  'lr': 2e-4,                                          # learning rate used during training
}

# create the Estimator
huggingface_estimator = HuggingFace(
    entry_point          = 'run_clm.py',      # train script
    source_dir           = 'scripts',         # directory which includes all the files needed for training
    instance_type        = 'ml.g5.2xlarge', # instances type used for the training job
    instance_count       = 1,                 # the number of instances used for training
    base_job_name        = job_name,          # the name of the training job
    role                 = role,              # IAM role used in training job to access AWS resources, e.g. S3
    volume_size          = 300,               # the size of the EBS volume in GB
    transformers_version = '4.26',            # the transformers version used in the training job
    pytorch_version      = '1.13',            # the pytorch_version version used in the training job
    py_version           = 'py39',            # the python version used in the training job
    hyperparameters      =  hyperparameters
)

You can now start your training job using the .fit() method and passing the S3 path to the training script:

# define a data input dictionary with our uploaded s3 uris
data = {'training': training_input_path}

# starting the train job with our uploaded datasets as inputs
huggingface_estimator.fit(data, wait=True)

Using LoRA and quantization makes fine-tuning BLOOMZ-7B to our task affordable and efficient with SageMaker. When using SageMaker training jobs, you only pay for GPUs for the duration of model training. In our example, the SageMaker training job took 20,632 seconds, which is about 5.7 hours. The ml.g5.2xlarge instance we used costs $1.515 per hour for on-demand usage. As a result, the total cost for training our fine-tuned BLOOMZ-7B model was only $8.63. Comparatively, full fine-tuning of the model’s 7 billion weights would cost an estimated $600, or 6,900% more per training run, assuming linear GPU scaling on the original computing configuration outlined in the Hugging Face model card. In practice, this would further vary depending upon your training strategy, instance selection, and instance pricing.

We could also further reduce our training costs by using SageMaker managed Spot Instances. However, there is a possibility this would result in the total training time increasing due to Spot Instance interruptions. See Amazon SageMaker Pricing for instance pricing details.

Deploy the model to a SageMaker endpoint for inference

With LoRA, you previously adapted a smaller set of weights to your new task. You need a way to combine these task-specific weights with the pre-trained weights of the original model. In the run_clm.py script, the PEFT library merge_and_unload() method takes care of merging the base BLOOMZ-7B model with the updated adapter weights fine-tuned to your task to make them easier to deploy without introducing any inference latency compared to the original model.

In this section, we go through the steps to create a SageMaker model from the fine-tuned model artifact and deploy it to a SageMaker endpoint for inference. First, you can create a Hugging Face model using your new fine-tuned model artifact for deployment to a SageMaker endpoint. Because you previously trained the model with a SageMaker Hugging Face estimator, you can deploy the model immediately. You could instead upload the trained model to an S3 bucket and use them to create a model package later. See the following code:

from sagemaker.huggingface import HuggingFaceModel

# 1. create Hugging Face Model Class
huggingface_model = HuggingFaceModel(
   model_data=huggingface_estimator.model_data,
   #model_data="s3://hf-sagemaker-inference/model.tar.gz",  # Change to your model path
   role=role, 
   transformers_version="4.26", 
   pytorch_version="1.13", 
   py_version="py39",
   model_server_workers=1
)

As with any SageMaker estimator, you can deploy the model using the deploy() method from the Hugging Face estimator object, passing in the desired number and type of instances. In this example, we use the same G5 instance type equipped with a single NVIDIA A10g GPU that the model was fine-tuned on in the previous step:

# 2. deploy model to SageMaker Inference
predictor = huggingface_model.deploy(
   initial_instance_count=1,
   instance_type= "ml.g5.4xlarge"
)

It may take 5–10 minutes for the SageMaker endpoint to bring your instance online and download your model in order to be ready to accept inference requests.

When the endpoint is running, you can test it by sending a sample dialog from the dataset test split. First load the test split using the Hugging Face Datasets library. Next, select a random integer for index slicing a single test sample from the dataset array. Using string formatting, combine the test sample with a prompt template into a structured input to guide our model’s response. This structured input can then be combined with additional model input parameters into a formatted sample JSON payload. Finally, invoke the SageMaker endpoint with the formatted sample and print the model’s output summarizing the sample dialog. See the following code:

from random import randint
from datasets import load_dataset

# 1. Load dataset from the hub
test_dataset = load_dataset("samsum", split="test")

# 2. select a random test sample
sample = test_dataset[randint(0,len(test_dataset))]

# 3. format the sample
prompt_template = f"Summarize the chat dialogue:n{{dialogue}}n---nSummary:n"

fomatted_sample = {
  "inputs": prompt_template.format(dialogue=sample["dialogue"]),
  "parameters": {
    "do_sample": True, # sample output predicted probabilities
    "top_p": 0.9, # sampling technique Fan et. al (2018)
    "temperature": 0.1, # increasing the likelihood of high probability words and decreasing the likelihood of low probability words
    "max_new_tokens": 100, # 
  }
}

# 4. Invoke the SageMaker endpoint with the formatted sample
res = predictor.predict(fomatted_sample)


# 5. Print the model output
print(res[0]["generated_text"].split("Summary:")[-1])
# Sample model output: Kirsten and Alex are going bowling this Friday at 7 pm. They will meet up and then go together.

print(sample["summary"])
# Sample model input: Kirsten reminds Alex that the youth group meets this Friday at 7 pm to go bowling.

Clean up

Now that you’ve tested your model, make sure that you clean up the associated SageMaker resources to prevent continued charges:

predictor.delete_model()
predictor.delete_endpoint()

Summary

In this post, you used the Hugging Face Transformer, PEFT, and the bitsandbytes libraries with SageMaker to fine-tune a BloomZ large language model on a single GPU for $8 and then deployed the model to a SageMaker endpoint for inference on a test sample. SageMaker offers multiple ways to use Hugging Face models; for more examples, check out the AWS Samples GitHub.

To continue using SageMaker to fine-tune foundation models, try out some of the techniques in the post Architect personalized generative AI SaaS applications on Amazon SageMaker. We also encourage you to learn more about Amazon Generative AI capabilities by exploring JumpStart, Amazon Titan models, and Amazon Bedrock.

About the Authors

Philipp Schmid is a Technical Lead at Hugging Face with the mission to democratize good machine learning through open source and open science. Philipp is passionate about productionizing cutting-edge and generative AI machine learning models. He loves to share his knowledge on AI and NLP at various meetups such as Data Science on AWS, and on his technical blog.

Robert Fisher is a Sr. Solutions Architect for Healthcare and Life Sciences customers. He works closely with customers to understand how AWS can help them solve problems, especially in the AI/ML space. Robert has many years of experience in software engineering across a range of industry verticals including medical devices, fintech, and consumer-facing applications.

Doug Kelly is an AWS Sr. Solutions Architect that serves as a trusted technical advisor to top machine learning startups in verticals ranging from machine learning platforms, autonomous vehicles, to precision agriculture. He is member of the AWS ML technical field community where he specializes in supporting customers with MLOps and ML inference workloads.

Read More

This post is co-written with Philipp Schmid and Jeff Boudier from Hugging Face.

Today, as part of Amazon Web Services’ partnership with Hugging Face, we are excited to announce the release of a new Hugging Face Deep Learning Container (DLC) for inference with Large Language Models (LLMs). This new Hugging Face LLM DLC is powered by Text Generation Inference (TGI), an open source, purpose-built solution for deploying and serving Large Language Models. TGI enables high-performance text generation using Tensor Parallelism and dynamic batching for the most popular open-source LLMs, including StarCoder, BLOOM, GPT-NeoX, StableLM, Llama, and T5.

Large Language Models are growing in popularity but can be difficult to deploy

LLMs have emerged as the leading edge of artificial intelligence, captivating developers and enthusiasts alike with their ability to comprehend and generate human-like text across diverse domains. These powerful models, such as those based on the GPT and T5 architectures, have experienced an unprecedented surge in popularity for a broad set of applications, including language understanding, conversational experiences, and automated writing assistance. As a result, companies across industries are seizing the opportunity to unlock their potential and offer new LLM-powered experiences in their applications.

Hosting LLMs at scale presents a unique set of complex engineering challenges. To provide an ideal user experience, an LLM hosting service should provide adequate response times while scaling to a large number of concurrent users. Given the high resource requirements of large models, general-purpose inference frameworks may not provide the optimizations required to maximize the utilization of available resources and provide the best possible performance.

Some of these optimizations include:

• Tensor parallelism to distribute the computation across multiple accelerators
• Model quantization to reduce the memory footprint of the model
• Dynamic batching of inference requests to improve throughput, and many others.

The Hugging Face LLM DLC provides these optimizations out of the box and makes it easier to host LLM models at scale.

Hugging Face’s Text Generation Inference simplifies LLM deployment

TGI is an open source, purpose-built solution for deploying Large Language Models (LLMs). It incorporates optimizations including tensor parallelism for faster multi-GPU inference, dynamic batching to boost overall throughput, and optimized transformers code using flash-attention for popular model architectures including BLOOM, T5, GPT-NeoX, StarCoder, and LLaMa.

With the new Hugging Face LLM Inference DLCs on Amazon SageMaker, AWS customers can benefit from the same technologies that power highly concurrent, low latency LLM experiences like HuggingChat, OpenAssistant, and Inference API for LLM models on the Hugging Face Hub, while enjoying SageMaker’s managed service capabilities, such as autoscaling, health checks, and model monitoring.

Get started with TGI on SageMaker Hosting

Let’s walk through a code example that deploys a GPT NeoX 20B parameter model on a SageMaker Endpoint. You can find our complete example notebook here.

First, make sure that the latest version of SageMaker SDK is installed:

%pip install sagemaker>=2.161.0

Then, we import the SageMaker Python SDK and instantiate a sagemaker_session to find the current region and execution role.

import sagemaker
from sagemaker.huggingface import HuggingFaceModel, get_huggingface_llm_image_uri
import time

sagemaker_session = sagemaker.Session()
region = sagemaker_session.boto_region_name
role = sagemaker.get_execution_role()

Next we retrieve the LLM image URI. We use the helper function get_huggingface_llm_image_uri() to generate the appropriate image URI for the Hugging Face Large Language Model (LLM) inference. The function takes a required parameter backend and several optional parameters. The backend specifies the type of backend to use for the model, the values can be “lmi” and “huggingface”.  “lmi” stands for SageMaker Large Model Inference backend and “huggingface” refers to using Hugging Face TGI backend that is used in this tutorial.

image_uri = get_huggingface_llm_image_uri(
  backend="huggingface", # or lmi
  region=region
)

Now that we have the image uri, the next step is to configure the model object. We specify a unique name, the image_uri for the managed TGI container, and the execution role for the endpoint. Additionally, we specify a number of environment variables including the HF_MODEL_ID which corresponds to the model from the HuggingFace Hub that will be deployed, and the HF_TASK which configures the inference task to be performed by the model.

You should also define SM_NUM_GPUS, which specifies the tensor parallelism degree of the model. Tensor parallelism can be used to split the model across multiple GPUs, which is necessary when working with LLMs that are too big for a single GPU. To learn more about tensor parallelism with inference, see our previous blog post. Here, you should set SM_NUM_GPUS to the number of available GPUs on your selected instance type. For example, in this tutorial, we set SM_NUM_GPUS to 4 because our selected instance type ml.g4dn.12xlarge has 4 available GPUs.

Note that you can optionally reduce the memory and computational footprint of the model by setting the HF_MODEL_QUANTIZE environment variable to “true”, but this lower weight precision could affect the quality of the output for some models.

model_name = "gpt-neox-20b-" + time.strftime("%Y-%m-%d-%H-%M-%S", time.gmtime())

hub = {
    'HF_MODEL_ID':'EleutherAI/gpt-neox-20b',
    'HF_TASK':'text-generation',
    'SM_NUM_GPUS':'4',
    'HF_MODEL_QUANTIZE':'true'
}

model = HuggingFaceModel(
    name=model_name,
    env=hub,
    role=role,
    image_uri=image_uri
)

Next, we invoke the deploy method to deploy the model.

predictor = model.deploy(
  initial_instance_count=1,
  instance_type="ml.g4dn.12xlarge",
  endpoint_name=model_name
)

Once the model is deployed, we can invoke it to generate text. We pass an input prompt and run the predict method to generate a text response from the LLM running in the TGI container.

input_data = {
  "inputs": "The diamondback terrapin was the first reptile to",
  "parameters": {
    "do_sample": True,
    "max_new_tokens": 100,
    "temperature": 0.7,
    "watermark": True
  }
}

predictor.predict(input_data)

We receive the following auto-generated text response:.

[{'generated_text': 'The diamondback terrapin was the first reptile to make the list, followed by the American alligator, the American crocodile, and the American box turtle. The polecat, a ferret-like animal, and the skunk rounded out the list, both having gained their slots because they have proven to be particularly dangerous to humans.nnCalifornians also seemed to appreciate the new list, judging by the comments left after the election.nn“This is fantastic,” one commenter declared.nn“California is a very'}]

To mitigate the risk of potential exploitation of Generative AI capabilities by automated bots, the response is watermarked. Such watermarked responses can be easily detected by algorithms, promoting the responsible use of Generative AI.

Once we are done experimenting, we delete the endpoint and the model resources.

predictor.delete_model()
predictor.delete_endpoint()

Conclusion and next steps

Deploying Large Language Models using Hugging Face’s Text Generation Inference and SageMaker Hosting is a straightforward solution for hosting open source models like GPT-NeoX, Flan-T5-XXL, StarCoder or LLaMa. The state of the art LLMs are deployed within the secure managed SageMaker environment, and AWS customers can benefit from Large Language Models while keeping full control over their implementation, and without sending their data over to a third-party API.

In the tutorial, we demonstrated the deployment of GPT-NeoX using the new Hugging Face LLM Inference DLC, leveraging the power of 4 GPUs on a SageMaker ml.g4dn.12xlarge  instance. With this approach, users can effortlessly harness the capabilities of state-of-the-art language models, enabling a wide range of applications and advancements in natural language processing.

As a next step, you can learn more about Hugging Face LLM Inference on SageMaker with the following resources:

• Example notebook from this tutorial
• Hugging Face TGI Repository
• Hugging Face TGI Launch Blog

About the authors

Philipp Schmid is a Technical Lead at Hugging Face with the mission to democratize good machine learning through open source and open science. Philipp is passionate about productionizing cutting-edge & generative AI machine learning models.

Jeff Boudier builds products at Hugging Face, the #1 open platform for AI builders. Previously Jeff was a co-founder of Stupeflix, acquired by GoPro, where he served as director of Product Management, Product Marketing,  Business Development and Corporate Development.

Robert Van Dusen is a Senior Product Manager with Amazon SageMaker. He leads deep learning model optimization for applications such as large model inference.

Qing Lan is a Software Development Engineer in AWS. He has been working on several challenging products in Amazon, including high performance ML inference solutions and high performance logging system. Qing’s team successfully launched the first Billion-parameter model in Amazon Advertising with very low latency required. Qing has in-depth knowledge on the infrastructure optimization and Deep Learning acceleration.

Simon Zamarin is an AI/ML Solutions Architect whose main focus is helping customers extract value from their data assets. In his spare time, Simon enjoys spending time with family, reading sci-fi, and working on various DIY house projects.

Xin Yang is a Software Development Engineer at AWS. She has been working on deploying and optimizing deep learning inference systems. Her work spans both the realms of real-time inference and scalable offline inference solutions. In her spare time, Xin enjoys reading and hiking.

Gagan Singh is a Senior Technical Account Manager at AWS helping digital native startups maximize business success. He helps customers with adoption and optimization of real-time, multi-model ML inferencing endpoints using Amazon SageMaker. In his spare time, Gagan enjoys trekking in the Himalayas and listening to music.

Read More