DeepNet

Parameters

3.2B

Published

Mar 1, 2022

DeepNet is an AI model developed by Microsoft Research (United States), first published in March 2022. It works in the language domain, on tasks such as language modeling, translation and language modeling/generation.

Epoch AI has no training-compute estimate for this model. The model has 3,200,000,000 parameters. It was trained on roughly 272.6B datapoints.

Access: Unreleased. Its weights are not openly released. The reference paper has 244 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Microsoft Research
Country of organization: United States
Domain: Language
Task: Language modeling, Translation, Language modeling/generation
Compute estimation method: Comparison with other models
Parameters: 3,200,000,000
Dataset size: 272.6B
Numerical format: FP16
Model accessibility: Unreleased
Open weights: No
Citations: 244
Epoch confidence: Confident