Hymba

Training compute

1.4×10²² FLOP

Parameters

1.5B

Published

Nov 22, 2024

Hymba is an AI model developed by NVIDIA (United States), first published in November 2024. It works in the language domain, on tasks such as language modeling/generation and question answering.

Training it took an estimated 1.4×10²² FLOP of compute (estimation method: operation counting). The model has 1,500,000,000 parameters. It was trained on roughly 1.5T datapoints. Training ran on 128 NVIDIA A100.

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

Full record

Organization: NVIDIA
Country of organization: United States
Domain: Language
Task: Language modeling/generation, Question answering
Training compute: 1.4×10²² FLOP
Compute estimation method: Operation counting
Parameters: 1,500,000,000
Dataset size: 1.5T
Training hardware: NVIDIA A100
Chips used: 128
Training power draw: 100.7 kW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident