Big-Little Net (vision)

Training compute

6.3×10¹⁹ FLOP

Parameters

77.4M

Published

Jul 10, 2018

Big-Little Net (vision) is an AI model developed by IBM (United States), first published in July 2018. It works in the vision domain, on tasks such as object recognition.

Training it took an estimated 6.3×10¹⁹ FLOP of compute (estimation method: operation counting). The model has 77,360,000 parameters. It was trained on roughly 1.3M datapoints.

Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 101 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: IBM
Country of organization: United States
Domain: Vision
Task: Object recognition
Training compute: 6.3×10¹⁹ FLOP
Compute estimation method: Operation counting
Parameters: 77,360,000
Dataset size: 1.3M
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Citations: 101
Epoch confidence: Confident