wav2vec2-xls-r-300m-bengali-commonvoice / preprocessor_config.json
shahruk10's picture
feat: add best model iteration from second run and language models
62b7482
{
"do_normalize": true,
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
"feature_size": 1,
"padding_side": "right",
"padding_value": 0.0,
"processor_class": "Wav2Vec2ProcessorWithLM",
"return_attention_mask": true,
"sampling_rate": 16000
}