Building A Faster & Accurate COVID Search Engine with Transformers🤗

This article is a step by step guide to build a faster and accurate COVID Semantic Search Engine using HuggingFace Transformers🤗. In this article, we will build a search engine, which will not only retrieve and rank the articles based on the query but also give us the response, along with a 1000 words context around the response

Fine Tuning XLNet Model for Text Classification

In this article, we will see how to fine tune a XLNet model on custom data, for text classification using Transformers🤗. XLNet is powerful! It beats BERT and its other variants in 20 different tasks. In simple words - XLNet is a generalized autoregressive model. An Autoregressive model is a model which uses the context word to predict the next word. So, the next token is dependent on all previous tokens. XLNET is generalized because it captures bi-directional context by means of a mechanism called permutation language modeling. It integrates the idea of auto-regressive models and bi-directional context modeling, yet overcoming the disadvantages of BERT and thus outperforming BERT on 20 tasks, often by a large margin in tasks such as question answering, natural language inference, sentiment analysis, and document ranking. In this article, we will take a pretrained `XLNet` model and fine tune it on our dataset.

Building Question Answering Model at Scale using 🤗Transformers

In this article, you will learn how to fetch contextual answers in a huge corpus of documents using Transformers🤗. We will build a neural question and answering system using transformers models (`RoBERTa`). This approach is capable to perform Q&A across millions of documents in few seconds.

Training a T5 Transformer Model - Generating Titles from ArXiv Paper's Abstracts using 🤗Transformers

In this article, you will learn how to train a `T5 model` for text generation - to generate title given a research paper's abstract or summary using Transformers🤗. For this tutorial, We will take research paper's abstract or brief summary as our input text and its corrosponding paper's title as output text and feed it to a `T5 model` to train. Once the model is trained, it will be able to generate the paper's title based on the abstract.