transferability and ensemble studies In the code, three models are included: the Deepspeech model, the Lingvo model, and Kaldi's aspire model In the code, three models are included: the Deepspeech model, the Lingvo model, and Kaldi's aspire model. If you want to know more, please refer to the following links:
https://github.com/mozilla/DeepSpeech
https://github.com/tensorflow/lingvo
https://github.com/kaldi-asr/kaldi
Among these files, the "ensembleattack.py" file is the program launcher, you can change the corresponding 'input_audio_path' and 'target' and then run.