1. What’s BLOOM?
BLOOM is an autoregressive LLM model developed by BigScience and published in Hugging Face which essentially works with a user prompt, similar to other models such as GPT-4. However, unlike the latter, BLOOM is free and open source thanks to the contribution of 1,000 researchers from more than 70 countries and trained on the Jean Zay public supercomputer, in the south of Paris between 2021 and 2022.
2. What’s the BLOOM function?
The usefulness of BLOOM lies in the meaning of its name “BigScience Large Open-science Open-access Multilingual Language Model” being an LLM created by collaboration of researchers to deliver an LLM that was free and accessible. This model works like any other of its type, in which the user uses a prompt to ask and access the model information, also having support in multiple languages, however it does not have a visual interface from which to interact, since this interaction is done by code.
User applications of BLOOM can range from something as simple as using it to study new languages, to automating processes similar to what would be done with other LLMs using Langchain as an intermediary tool to achieve this. It could even be used for research by helping to write the text and offering both technical and non-technical information on the topic you ask.
3. BLOOM Features.
- LLM prompt model similar to GPT-4.
- Support in 46 languages.
- Generate code in 13 programming languages.
- It was trained with 176 billion parameters.
- Completely free and open source.
- The user prompt is done by code.