BLAZE (Build Language Applications Easily) is an open-source project by Cisco for creating customizable and composable Natural Language Understanding (NLU) inference pipelines. It is a flexible, standardized, no-code platform that allows users to easily assemble, modify, and deploy various NLP models, datasets, and components.
BLAZE works by abstracting out the functionality of different stages of the NLP pipeline, such as tokenization, embedding, classification, summarization, generation, etc. These stages are represented as flexible blocks that can be combined in various ways to create new recipes of NLP pipelines. Users can also add and arrange their own custom blocks to extend the functionality of BLAZE.
BLAZE is used for building and testing NLP applications that can leverage the power of large language models (LLMs) and other state-of-the-art techniques. Some examples of NLP applications that can be built with BLAZE are:
- Meeting assistants: Users can create a Webex embedded app that uses BLAZE to provide LLM-powered insights to meeting attendees, such as action items, summaries, questions, etc.
- Chatbots: Users can create a Webex chatbot that uses BLAZE to interact with users using natural language and provide relevant information or services.
- Content creation: Users can use BLAZE to generate imaginative and innovative content such as poems, stories, code, essays, songs, celebrity parodies, etc., using LLMs and other models.