XGen-7B

Training compute

8×10²² FLOP

Parameters

6.7B

Published

Sep 7, 2023

XGen-7B is an AI model developed by Salesforce (United States), first published in September 2023. It works in the language domain, on tasks such as language generation.

Training it took an estimated 8×10²² FLOP of compute (estimation method: hardware,operation counting). The model has 6,700,000,000 parameters. It was trained on roughly 1.4T datapoints. Training ran on Google TPU v4.

Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 17 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Salesforce
Country of organization: United States
Domain: Language
Task: Language generation
Training compute: 8×10²² FLOP
Compute estimation method: Hardware, Operation counting
Parameters: 6,700,000,000
Dataset size: 1.4T
Training hardware: Google TPU v4
Chip-hours: 270.3K
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Citations: 17
Epoch confidence: Confident

Benchmark results

01

Epoch Capabilities Index

ECI Score

93