AFP+FPI (PTB)

Training compute

2.2×10¹⁴ FLOP

Parameters

2M

Published

Jun 4, 2021

AFP+FPI (PTB) is an AI model developed by University of Sheffield (United Kingdom), first published in June 2021. It works in the language domain, on tasks such as language modeling.

Training it took an estimated 2.2×10¹⁴ FLOP of compute (estimation method: operation counting). The model has 2,040,000 parameters. It was trained on roughly 912.3K datapoints.

Access: Unreleased. Its weights are not openly released. The reference paper has 4 citations. Epoch AI rates the confidence of this record as likely.

Full record

Organization: University of Sheffield
Country of organization: United Kingdom
Domain: Language
Task: Language modeling
Training compute: 2.2×10¹⁴ FLOP
Compute estimation method: Operation counting
Parameters: 2,040,000
Dataset size: 912.3K
Model accessibility: Unreleased
Open weights: No
Citations: 4
Epoch confidence: Likely