StableLM-3B-4E1T

Training compute

6.2×10²² FLOP

Parameters

2.8B

Published

Sep 29, 2023

StableLM-3B-4E1T is an AI model developed by Stability AI (United Kingdom), first published in September 2023. It works in the language domain, on tasks such as language generation.

Training it took an estimated 6.2×10²² FLOP of compute (estimation method: hardware). The model has 2,795,443,200 parameters. It was trained on roughly 1T datapoints. Training ran on NVIDIA A100 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: Stability AI
Country of organization: United Kingdom
Domain: Language
Task: Language generation
Training compute: 6.2×10²² FLOP
Compute estimation method: Hardware
Parameters: 2,795,443,200
Dataset size: 1T
Training hardware: NVIDIA A100
Training time: 720 h
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident