Shortcuts

Model Hub Transformers

Overview

The Huggingface transformers library is the de-facto library for natural language processing (NLP) models. It provides pretrained weights for leading NLP models and allows users to easily use these pretrained models for the most common NLP tasks like language modeling, text classification, and question answering among others.

model-hub makes it easy for users to train transformer models in Determined while keeping the developer experience as close as possible to what it’s like working directly with transformers. Our library serves as an alternative to HuggingFace’s Trainer Class and provides access to all of Determined’s benefits including:

Note

Model Hub Transformers is similar to the no_trainer version of transformers examples in that we allow users to have more control over the training and evaluation routines if desired.

Given the benefits above, we think this library will be particularly useful to you if any of the following apply:

  • You are an Determined user that wants to get started quickly with transformers.

  • You are a transformers user that wants to easily run more advanced workflows like multi-node distributed training and advanced hyperparameter search.

  • You are a transformers user looking for a single platform to manage experiments, handle checkpoints with automated fault tolerance, and perform hyperparameter search/visualization.

Getting Started

The easiest way to use Model Hub Transformers is to start with an existing example Trial. Model Hub Transformers includes thoroughly tested implementations of all core transformers tasks.

Model Hub Transformers Trials are infinitely customizable. Learn about how to customize or build your own Trial by checking out the Model Hub Transformers Tutorial.

Limitations

While we strive to offer as many of the features supported in HuggingFace transformers, the following features are not currently supported:

  • Tensorflow version of transformers

  • Support for deepspeed and fairscale

  • Running on TPUs