from datasets import load_dataset
dataset = load_dataset('disco-eth/AIME-survey')
This survey dataset accompanies the AIME audio dataset.
The AIME Survey dataset consists of 15,600 pairwise audio comparisons rated by more than 2,500 human participants regarding the music quality and text-audio alignment of 12 state-of-the-art music generation models (as of July 2024). The comparisons were made between 10 second snippets of the audio tracks.… See the full description on the dataset page:
https://huggingface.co/datasets/disco-eth/AIME-survey.
