Natural Language Processing has created a new paradigm shift in both academia and industry today. With the cominig up of Transformers, the field has seen great growth and applications to fields which were previously unheard of. Google’s BERT has further accelerated the progress in the field. This paper can be considered as the backbone of modern NLP. Multi-Modal systems utilize data from 2 or more modalities to make decisions. Our goal with this library is to provide an efficient, fast and reliable way to load and train models on multi-modal data.
With Tochmm, our goal is three-fold: - To educate the user about Natural Language Processing and Multi-Modal Systems.
- Easy to understand implementations of State of the Multi-Modal Algorithms.
- Develop efficient pipelines of existing Algorithms.