| # pip install numpy==1.26.4 transformers==4.45.2 datasets==3.0.1 librosa==0.10.2.post1 jiwer==3.0.4 bitsandbytes==0.38.0.post1 accelerate==0.26.0 | |
| dataset_name = "dlxjj/common_voice_17_0" | |
| language = "Japanese" | |
| language_abbr = "ja" | |
| from datasets import load_dataset, DatasetDict | |
| common_voice = DatasetDict() | |
| common_voice["train"] = load_dataset(dataset_name, language_abbr, split="train", trust_remote_code=True) | |
| print(common_voice["train"][0]) |