Wav2vec2.0-xlsr-53 model is fine-tuned on commonvoice russian dataset