AFM-server

Training compute

4.3×10²⁴ FLOP

Published

Jul 29, 2024

AFM-server is an AI model developed by Apple (United States), first published in July 2024. It works in the language domain, on tasks such as language modeling/generation.

Training it took an estimated 4.3×10²⁴ FLOP of compute (estimation method: operation counting,hardware). It was trained on roughly 7.4T datapoints. Training ran on 8,192 Google TPU v4.

Access: Hosted access (no API). Its weights are not openly released. Epoch AI rates the confidence of this record as likely.

Full record

Organization: Apple
Country of organization: United States
Domain: Language
Task: Language modeling/generation
Training compute: 4.3×10²⁴ FLOP
Compute estimation method: Operation counting, Hardware
Dataset size: 7.4T
Training hardware: Google TPU v4
Chips used: 8,192
Training power draw: 5.5 MW
Numerical format: BF16
Model accessibility: Hosted access (no API)
Open weights: No
Epoch confidence: Likely