🤗 Optimum provides an integration with
Better Transformer, a fast path of standard PyTorch Transformer APIs to benefit from interesting speedups on CPU & GPU through sparsity and fused kernels. For now, it supports Transformer encoders, basically fast path of
support for decoders and training path is coming soon.
Since its 1.13 version, PyTorch released the stable version of a fast path for its standard Transformer APIs that provides out of the box performance improvements for transformer-based models. You can benefit from interesting speedup on most consumer-type devices, including CPUs, older and newer versions of NIVIDIA GPUs. You can now use this feature in 🤗 Optimum together with Transformers and use it for major models in the Hugging Face ecosystem.
The list of supported model below:
Note that for encoder-decoder models, only the encoder part is supported by PyTorch’s BetterTransformer for now.
Let us know by opening an issue in 🤗 Optimum if you want more models to be supported, or check out the contribution guideline if you want to add it by yourself!
In order to use the
BetterTransformer API just run the following commands:
from transformers import AutoModelForSequenceClassification from optimum.bettertransformer import BetterTransformer model_hf = AutoModelForSequenceClassification.from_pretrained("bert-base-cased") model = BetterTransformer.transform(model_hf, keep_original_model=True)
You can leave
keep_original_model=False in case you want to overwrite the current model with its
More details on
tutorials section to deeply understand how to use it, or check the Google colab demo!