File size: 703 Bytes
affcd23 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 |
import logging
from pathlib import Path
import torchaudio
from hw_asr.base.base_dataset import BaseDataset
logger = logging.getLogger(__name__)
class CustomAudioDataset(BaseDataset):
def __init__(self, data, *args, **kwargs):
index = data
for entry in data:
assert "path" in entry
assert Path(entry["path"]).exists(), f"Path {entry['path']} doesn't exist"
entry["path"] = str(Path(entry["path"]).absolute().resolve())
entry["text"] = entry.get("text", "")
t_info = torchaudio.info(entry["path"])
entry["audio_len"] = t_info.num_frames / t_info.sample_rate
super().__init__(index, *args, **kwargs)
|