AudioLM

Training compute

3.9×10¹⁸ FLOP

Parameters

1.5B

Published

Jul 26, 2023

AudioLM is an AI model developed by Google Research (United States), first published in July 2023. It works in the audio domain, on tasks such as audio generation.

Training it took an estimated 3.9×10¹⁸ FLOP of compute (estimation method: operation counting). The model has 1,500,000,000 parameters. It was trained on roughly 135B datapoints. Training ran on Google TPU v4.

Access: Unreleased. Its weights are not openly released. The reference paper has 917 citations. Epoch AI rates the confidence of this record as speculative.

Full record

Organization: Google Research
Country of organization: United States
Domain: Audio
Task: Audio generation
Training compute: 3.9×10¹⁸ FLOP
Compute estimation method: Operation counting
Parameters: 1,500,000,000
Dataset size: 135B
Training hardware: Google TPU v4
Model accessibility: Unreleased
Open weights: No
Citations: 917
Epoch confidence: Speculative