RedPajama-INCITE-7B-Base

Training compute

4.1×10²² FLOP

Parameters

6.9B

Published

Jun 6, 2023

RedPajama-INCITE-7B-Base is an AI model developed by Together (United States), first published in June 2023. It works in the language domain, on tasks such as chat.

Training it took an estimated 4.1×10²² FLOP of compute (estimation method: operation counting). The model has 6,900,000,000 parameters. It was trained on roughly 1T datapoints. Training ran on 3,072 NVIDIA V100.

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

Full record

Organization: Together
Country of organization: United States
Domain: Language
Task: Chat
Training compute: 4.1×10²² FLOP
Compute estimation method: Operation counting
Parameters: 6,900,000,000
Dataset size: 1T
Training hardware: NVIDIA V100
Chips used: 3,072
Training power draw: 1.8 MW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident

Benchmark results

01	PIQA	Score	0.77
02	HellaSwag	Overall accuracy	0.7
03	BoolQ	Score	0.69
04	WinoGrande	Accuracy	0.64
05	OpenBookQA	Accuracy	0.4
06	ARC AI2	Challenge score	0.39
07	MMLU	EM	0.26