Press Release | Grenoble, France | September 16th, 2021
NLP Cloud API
Launched in 2020, Grenoble-based NLP Cloud provides high performance Natural Language Processing (NLP) models based on spaCy and Hugging Face Transformers, ready for production. The company announces its compatibility with GPT-J, a 6-billion-parameter NLP AI model, competing with GPT-3. GPT-3’s source code belongs to Microsoft. It is available to selected customers only, and at a premium price. On the other hand, GPT-J (and GPT-Neo, its predecessor) is an open-source model, released by a team of researchers from EleutherAI.
GPT-J is the most advanced open-source NLP model as of this writing.
GPT-3 and GPT-J have recently disrupted the NLP world. Thanks to their huge sizes, these AI models are both very good at generating fluent human language, and at addressing efficiently most of the modern NLP use cases. Thanks to GPT-J, developers can perform various state-of-the-art operations such as chatbots, data extraction from unstructured documents (like web pages, emails, contracts, financial statements,…), code generation, sentiment analysis, or marketing content creation, in many languages. New use cases are discovered almost everyday.
Running GPT-J efficiently in production is a technical challenge. The NLP Cloud team has been working for the last three months on the GPT-J integration to offer its clients a production-ready solution.
“At NLP Cloud, we are happy to propose an open-source alternative to the OpenAI-Microsoft monopoly, as we think that anyone should be able to understand – if s/he wants to – how an AI model is working under the hood” says Julien Salinas, founder and CEO at NLP Cloud.
Advanced NLP models pose new ethical challenges. In order to fight against potentially dangerous models and biases like racism and misogyny, NLP Cloud thinks that the only solution is for NLP models to remain open-source and continuously audited by the community.
Founded in 2020, NLP Cloud is an API that supports the interaction between computers and the human language. By mixing Natural Language Processing and DevOps, NLP Cloud helps app developers and data scientists worldwide add language processing intelligence to their apps, in production. The company is based in Grenoble, France, and supports all NLP use cases at scale, from sentiment-analysis and text classification to summarization and text generation. Its models are currently available in many languages, for global distribution.
Website - Linkedin - Twitter - GitHub
Logo - Julien Salinas' picture
Pauline Vettier, Press Officer, NLP Cloud
Email: [email protected]