Write With Transformer

Write With Transformer is a web application built by the Hugging Face team that serves as the official demonstration of the 🤗/transformers repository's text generation capabilities. It allows users to type text and receive neural network-generated continuations, providing a hands-on way to explore how different language models complete written passages.

The app offers several selectable models, including GPT-2, XLNet, GPT, DistilGPT-2, and Arxiv-NLP. Each model entry includes a description of its architecture and training approach, along with a "Start writing" option to begin generating text. DistilGPT-2 is presented as a distilled, lighter-weight version of GPT-2, while Arxiv-NLP is a GPT-2 model fine-tuned on a dataset of Arxiv papers focused on Natural Language Processing. The interface also includes a section for featured public documents and an invitation to contribute or suggest new model checkpoints through the 🤗/transformers repository.

The application is aimed at users interested in exploring text generation models and their capabilities. No pricing, trial, or edition information is provided on the page.