espnet2.train.preprocessor.Qwen2AudioPreprocessor

source

class espnet2.train.preprocessor.Qwen2AudioPreprocessor(sampling_rate: int = 16000)

Bases: AbsPreprocessor

Preprocessor specifically for Qwen2Audio models

Initialize the Qwen2AudioPreprocessor.

This method sets up the tokenizer for Qwen2Audio models and defines the default sampling rate for audio processing.