This dataset contains tokenized audio from lewtun/music_genres using SemantiCodec for performing experiments on AR music generation
The following script is used for tokenization
from datasets import load_dataset, Dataset, DatasetDict
model_id = ""
repo_name = ""
user_name = ""
token = ""
cache_dir = "cache"
vocab_size = 4096
dataset = load_dataset(model_id, cache_dir=cache_dir, trust_remote_code=True)
from semanticodec import SemantiCodec
semanticodec = SemantiCodec(token_rate=100… See the full description on the dataset page:
https://huggingface.co/datasets/darshanmakwana/music_genre_tokenized.
