PLaMo-13B

Training compute

1.2×10²³ FLOP

Parameters

13B

Published

Sep 28, 2023

PLaMo-13B is an AI model developed by Preferred Networks Inc (Japan), first published in September 2023. It works in the language domain, on tasks such as language modeling/generation, chat and question answering.

Training it took an estimated 1.2×10²³ FLOP of compute (estimation method: operation counting). The model has 13,000,000,000 parameters. It was trained on roughly 1.5T datapoints. Training ran on 480 NVIDIA A100 SXM4 40 GB for about 720 hours.

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

Full record

Organization: Preferred Networks Inc
Country of organization: Japan
Domain: Language
Task: Language modeling/generation, Chat, Question answering
Training compute: 1.2×10²³ FLOP
Compute estimation method: Operation counting
Parameters: 13,000,000,000
Dataset size: 1.5T
Training hardware: NVIDIA A100 SXM4 40 GB
Chips used: 480
Training time: 720 h
Chip-hours: 345.6K
Training power draw: 381.3 kW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident