RNNLM + Dynamic KL Regularization (WT2)

Training compute

2.1×10¹⁶ FLOP

Parameters

87.6M

Published

Apr 27, 2018

RNNLM + Dynamic KL Regularization (WT2) is an AI model developed by Northwestern University (United States), first published in April 2018. It works in the language domain, on tasks such as language modeling.

Training it took an estimated 2.1×10¹⁶ FLOP of compute (estimation method: operation counting). The model has 87,600,000 parameters. It was trained on roughly 2M datapoints.

Access: Unreleased. Its weights are not openly released. The reference paper has 9 citations. Epoch AI rates the confidence of this record as likely.

Full record

Organization: Northwestern University
Country of organization: United States
Domain: Language
Task: Language modeling
Training compute: 2.1×10¹⁶ FLOP
Compute estimation method: Operation counting
Parameters: 87,600,000
Dataset size: 2M
Model accessibility: Unreleased
Open weights: No
Citations: 9
Epoch confidence: Likely