OpenELM-3B

Training compute

3.4×10²² FLOP

Parameters

3B

Published

May 2, 2024

OpenELM-3B is an AI model developed by Apple (United States), first published in May 2024. It works in the language domain, on tasks such as language modeling/generation, code generation and question answering.

Training it took an estimated 3.4×10²² FLOP of compute (estimation method: operation counting,hardware). The model has 3,040,000,000 parameters. Training ran on 128 NVIDIA H100 SXM5 80GB for about 288 hours.

Access: Open weights (non-commercial). Its weights are openly available. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Apple
Country of organization: United States
Domain: Language
Task: Language modeling/generation, Code generation, Question answering
Training compute: 3.4×10²² FLOP
Compute estimation method: Operation counting, Hardware
Parameters: 3,040,000,000
Training hardware: NVIDIA H100 SXM5 80GB
Chips used: 128
Training time: 288 h
Training power draw: 177.1 kW
Model accessibility: Open weights (non-commercial)
Open weights: Yes
Epoch confidence: Confident