As of the time of writing this piece, state-of-the-art results on NLP and NLU tasks are obtained with Transformer models. There is a trend of performance improvement as models become deeper and larger, GPT 3 comes to mind. Training small versions of such models from scratch takes a significant amount of time, even with GPU. This problem can be solved via pre-training when a model is trained on a large text corpus using a high-performance cluster. Later it can be fine-tuned for a specific task in a much shorter amount of time. During fine tuning stage, additional layers can be added to the model for specific tasks, which can be different from those for which the model was initially trained. This technique is related to transfer learning, a concept applied to areas of machine learning beyond NLP (see here and here for a quick intro).

In this post, I would like to share my experience of fine-tuning BERT and RoBERTa, available from the transformers library by Hugging Face, for a document classification task. Both models share a transformer architecture, which consists of at least two distinct blocks — encoder and decoder. Both encoder and decoder consist of multiple layers based around Attention mechanism. Encoder processed the input token sequence into a vector of floating point numbers — a hidden state, which is picked up by the decoder. It is the hidden state that encompasses the information content of the input sequence. This enables to represent an entire sequence of tokens with a single dense vector of float point numbers. Two texts or documents, which have similar meaning are represented by closely aligned vectors. Comparing vectors using a metric of choice, for example, cosine similarity, enables to quantify the similarity of original text pieces.

#machine-learning #nlp #data-science #text-classification #pytorch