train.sample.tsv: 
    - A sample of the dataframe used in the training process
str_ast_dataset.py: 
    - It contains classes and functions for the dataset used during the STR augmentation
speech_to_text.py: 
    - Slight modification on the original speech_to_text task for switching between the new dataset class and the original dataset class
speech_to_text_dataset.py: 
    - Slight modification on the class S2TDataConfig. Add some new arguments as inputs in the config files


