Dev gzf decoding (#1695)

zhifu gao

2024-05-06 00d0df3a1018c63ec8c5d13e611f53c564c0a7e2

 funasr/datasets/audio_datasets/update_jsonl.py

@@ -46,6 +46,7 @@
    data = json.loads(line.strip())

    wav_path = data["source"].replace("/cpfs01", "/cpfs_speech/data")
    if os.path.exists(wav_path):
    waveform, _ = librosa.load(wav_path, sr=16000)
    sample_num = len(waveform)
    source_len = int(sample_num / 16000 * 1000 / 10)