Generative BST

Training compute

1.4×10²² FLOP

Parameters

9.4B

Published

Mar 5, 2021

Generative BST is an AI model developed by Facebook AI Research (United States and France), first published in March 2021. It works in the language domain, on tasks such as language modeling/generation, chat and question answering.

Training it took an estimated 1.4×10²² FLOP of compute (estimation method: operation counting). The model has 9,431,810,048 parameters. It was trained on roughly 56.8B datapoints.

Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 1,123 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Facebook AI Research
Country of organization: United States, France
Domain: Language
Task: Language modeling/generation, Chat, Question answering
Training compute: 1.4×10²² FLOP
Compute estimation method: Operation counting
Parameters: 9,431,810,048
Dataset size: 56.8B
Numerical format: FP16
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Citations: 1,123
Epoch confidence: Confident