Software:GPT-J

From HandWiki
Short description: Open source artificial intelligence text generating language model developed by Eleuther AI
GPT-J
Logo
Logo
Developer(s)Eleuther AI
Initial releaseJune 9, 2021; 2 years ago (2021-06-09)
TypeLanguage model
LicenseOpen-source

GPT-J is an open source artificial intelligence language model developed by Eleuther AI.[1] GPT-J performs very similarly to OpenAI's GPT-3 on various zero-shot down-streaming tasks and can even outperform GPT-3.5 175 billion parameter dataset in code generation tasks.[2] The newest version, GPT-J-6B is a language model based on a 6 billion parameter data set called The Pile.[3] The Pile is an open-source 825 gigibyte language modelling data set that is split into 22 smaller datasets.[4] GPT-J is similar to ChatGPT in its ability, although it does not function as a chat bot, only as a text predictor.[5]

References

  1. ↑ Demo, GPT-3. "GPT-J | Discover AI use cases" (in en). https://gpt3demo.com/apps/gpt-j-6b. 
  2. ↑ "GPT-J-6B: An Introduction to the Largest Open Source GPT Model | Forefront" (in en). https://www.forefront.ai/blog-posts/gpt-j-6b-an-introduction-to-the-largest-open-sourced-gpt-model. 
  3. ↑ Wang, Ben (2023-02-28), Table of contents, https://github.com/kingoflolz/mesh-transformer-jax, retrieved 2023-02-28 
  4. ↑ "The Pile". https://pile.eleuther.ai/. 
  5. ↑ Mueller, Vincent (2022-01-25). "How you can use GPT-J" (in en). https://towardsdatascience.com/how-you-can-use-gpt-j-9c4299dd8526. 




Retrieved from "https://handwiki.org/wiki/index.php?title=Software:GPT-J&oldid=3673648"