Skip to content

🏎️ Accelerate training and inference of πŸ€— Transformers with easy to use hardware optimization tools

License

Notifications You must be signed in to change notification settings

fxmarty/optimum

Β 
Β 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

ONNX Runtime neural_compressor

Hugging Face Optimum

πŸ€— Optimum is an extension of πŸ€— Transformers, providing a set of performance optimization tools enabling maximum efficiency to train and run models on targeted hardware.

The AI ecosystem evolves quickly and more and more specialized hardware along with their own optimizations are emerging every day. As such, Optimum enables users to efficiently use any of these platforms with the same ease inherent to transformers.

Integration with Hardware Partners

πŸ€— Optimum aims at providing more diversity towards the kind of hardware users can target to train and finetune their models.

To achieve this, we are collaborating with the following hardware manufacturers in order to provide the best transformers integration:

Optimizing models towards inference

Along with supporting dedicated AI hardware for training, Optimum also provides inference optimizations towards various frameworks and platforms.

We currently support ONNX runtime along with Intel Neural Compressor (INC).

Features ONNX Runtime Intel Neural Compressor
Post-training Dynamic Quantization βœ”οΈ βœ”οΈ
Post-training Static Quantization βœ”οΈ βœ”οΈ
Quantization Aware Training (QAT) Stay tuned! ⭐ βœ”οΈ
Pruning N/A βœ”οΈ

Installation

πŸ€— Optimum can be installed using pip as follows:

python -m pip install optimum

If you'd like to use the accelerator-specific features of πŸ€— Optimum, you can install the required dependencies according to the table below:

Accelerator Installation
ONNX runtime python -m pip install optimum[onnxruntime]
Intel Neural Compressor (INC) python -m pip install optimum[intel]
Graphcore IPU python -m pip install optimum[graphcore]
Habana Gaudi Processor (HPU) python -m pip install optimum[habana]

If you'd like to play with the examples or need the bleeding edge of the code and can't wait for a new release, you can install the base library from source as follows:

python -m pip install git+https://github.com/huggingface/optimum.git

For the accelerator-specific features, you can install them by appending #egg=optimum[accelerator_type] to the pip command, e.g.

python -m pip install git+https://github.com/huggingface/optimum.git#egg=optimum[onnxruntime]

Quickstart

At its core, πŸ€— Optimum uses configuration objects to define parameters for optimization on different accelerators. These objects are then used to instantiate dedicated optimizers, quantizers, and pruners.

Quantization

For example, here's how you can apply dynamic quantization with ONNX Runtime:

from optimum.onnxruntime.configuration import AutoQuantizationConfig
from optimum.onnxruntime import ORTQuantizer

# The model we wish to quantize
model_checkpoint = "distilbert-base-uncased-finetuned-sst-2-english"
# The type of quantization to apply
qconfig = AutoQuantizationConfig.arm64(is_static=False, per_channel=False)
quantizer = ORTQuantizer.from_pretrained(model_checkpoint, feature="sequence-classification")

# Quantize the model!
quantizer.export(
    onnx_model_path="model.onnx",
    onnx_quantized_model_output_path="model-quantized.onnx",
    quantization_config=qconfig,
)

In this example, we've quantized a model from the Hugging Face Hub, but it could also be a path to a local model directory. The feature argument in the from_pretrained() method corresponds to the type of task that we wish to quantize the model for. The result from applying the export() method is a model-quantized.onnx file that can be used to run inference. Here's an example of how to load an ONNX Runtime model and generate predictions with it:

from functools import partial
from datasets import Dataset
from optimum.onnxruntime.model import ORTModel

# Load quantized model
ort_model = ORTModel("model-quantized.onnx", quantizer._onnx_config)
# Create a dataset or load one from the Hub
ds = Dataset.from_dict({"sentence": ["I love burritos!"]})
# Tokenize the inputs
def preprocess_fn(ex, tokenizer):
    return tokenizer(ex["sentence"])

tokenized_ds = ds.map(partial(preprocess_fn, tokenizer=quantizer.tokenizer))
ort_outputs = ort_model.evaluation_loop(tokenized_ds)
# Extract logits!
ort_outputs.predictions

Similarly, you can apply static quantization by simply setting is_static to True when instantiating the QuantizationConfig object:

qconfig = AutoQuantizationConfig.arm64(is_static=True, per_channel=False)

Static quantization relies on feeding batches of data through the model to estimate the activation quantization parameters ahead of inference time. To support this, πŸ€— Optimum allows you to provide a calibration dataset. The calibration dataset can be a simple Dataset object from the πŸ€— Datasets library, or any dataset that's hosted on the Hugging Face Hub. For this example, we'll pick the sst2 dataset that the model was originally trained on:

from optimum.onnxruntime.configuration import AutoCalibrationConfig

# Create the calibration dataset
calibration_dataset = quantizer.get_calibration_dataset(
    "glue",
    dataset_config_name="sst2",
    preprocess_function=partial(preprocess_fn, tokenizer=quantizer.tokenizer),
    num_samples=50,
    dataset_split="train",
)
# Create the calibration configuration containing the parameters related to calibration.
calibration_config = AutoCalibrationConfig.minmax(calibration_dataset)
# Perform the calibration step: computes the activations quantization ranges
ranges = quantizer.fit(
    dataset=calibration_dataset,
    calibration_config=calibration_config,
    onnx_model_path="model.onnx",
    operators_to_quantize=qconfig.operators_to_quantize,
)
# Quantize the same way we did for dynamic quantization!
quantizer.export(
    onnx_model_path="model.onnx",
    onnx_quantized_model_output_path="model-quantized.onnx",
    calibration_tensors_range=ranges,
    quantization_config=qconfig,
)

Graph optimization

Then let's take a look at applying graph optimizations techniques such as operator fusion and constant folding. As before, we load a configuration object, but this time by setting the optimization level instead of the quantization approach:

from optimum.onnxruntime.configuration import OptimizationConfig

# optimization_config=99 enables all available graph optimisations
optimization_config = OptimizationConfig(optimization_level=99)

Next, we load an optimizer to apply these optimisations to our model:

from optimum.onnxruntime import ORTOptimizer

optimizer = ORTOptimizer.from_pretrained(
    model_checkpoint,
    feature="sequence-classification",
)

# Export the optimized model
optimizer.export(
    onnx_model_path="model.onnx",
    onnx_optimized_model_output_path="model-optimized.onnx",
    optimization_config=optimization_config,
)

And that's it - the model is now optimized and ready for inference!

As you can see, the process is similar in each case:

  1. Define the optimization / quantization strategies via an OptimizationConfig / QuantizationConfig object
  2. Instantiate an ORTQuantizer or ORTOptimizer class
  3. Apply the export() method
  4. Run inference

Training

Besides supporting ONNX Runtime inference, πŸ€— Optimum also supports ONNX Runtime training, reducing the memory and computations needed during training. This can be achieved by using the class ORTTrainer, which possess a similar behavior than the Trainer of πŸ€— Transformers:

-from transformers import Trainer
+from optimum.onnxruntime import ORTTrainer

# Step 1: Create your ONNX Runtime Trainer
-trainer = Trainer(
+trainer = ORTTrainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
    eval_dataset=eval_dataset,
    compute_metrics=compute_metrics,
    tokenizer=tokenizer,
    data_collator=default_data_collator,
    feature="sequence-classification",
)

# Step 2: Use ONNX Runtime for training and evalution!πŸ€—
train_result = trainer.train()
eval_metrics = trainer.evaluate()

By replacing Trainer by ORTTrainer, you will be able to leverage ONNX Runtime for fine-tuning tasks.

Check out the examples directory for more sophisticated usage.

Happy optimizing πŸ€—!

About

🏎️ Accelerate training and inference of πŸ€— Transformers with easy to use hardware optimization tools

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 99.7%
  • Makefile 0.3%