Corrupt audio sample in "Telugu - train" split
#5
by
skesiraju
- opened
from datasets import load_dataset
indic_voices = load_dataset("ai4bharat/IndicVoices", "telugu")
indic_voices['train']['audio_filepath'][64251].get_all_samples()
Throws RuntimeError: Could not receive frame from decoder: Invalid data found when processing input
This is just for the sample at index 64251.
The metadata is accessible
indic_voices['train']['audio_filepath'][64251].metadata
AudioStreamMetadata:
duration_seconds_from_header: 1.792
begin_stream_seconds_from_header: 0.0
bit_rate: 0.0
codec: flac
stream_index: 0
sample_rate: 16000
num_channels: 1
sample_format: s16
Updated, please check
tahirjm
changed discussion status to
closed