ACF-WIDER

Training compute

7.6×10¹³ FLOP

Parameters

6.1K

Published

Jul 15, 2014

ACF-WIDER is an AI model developed by Chinese Academy of Sciences (China), first published in July 2014. It works in the vision domain, on tasks such as face detection.

Training it took an estimated 7.6×10¹³ FLOP of compute (estimation method: hardware). The model has 6,144 parameters. It was trained on roughly 144.4K datapoints.

Access: Unreleased. Its weights are not openly released. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Chinese Academy of Sciences
Country of organization: China
Domain: Vision
Task: Face detection
Training compute: 7.6×10¹³ FLOP
Compute estimation method: Hardware
Parameters: 6,144
Dataset size: 144.4K
Chips used: 1
Model accessibility: Unreleased
Open weights: No
Epoch confidence: Confident