Attend-Infer-Repeat

Training compute

6.4×10¹⁶ FLOP

Parameters

82.1M

Published

Aug 12, 2016

Attend-Infer-Repeat is an AI model developed by Google DeepMind (United States), first published in August 2016. It works in the vision domain, on tasks such as object recognition.

Training it took an estimated 6.4×10¹⁶ FLOP of compute (estimation method: hardware). The model has 82,130,304 parameters. Training ran on NVIDIA Quadro K4000 for about 48 hours.

The reference paper has 581 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Google DeepMind
Country of organization: United States
Domain: Vision
Task: Object recognition
Training compute: 6.4×10¹⁶ FLOP
Compute estimation method: Hardware
Parameters: 82,130,304
Training hardware: NVIDIA Quadro K4000
Training time: 48 h
Citations: 581
Epoch confidence: Confident