File size: 1,558 Bytes
30414af 4ebcdee 30414af |
1 2 3 4 5 6 7 8 9 |
# Noisy join mixed ASR
Problem with common ASR dataset, each audio sample is a monolanguage, but we can generate an audio sample with multilanguage using ASR and Force Alignment models. Notebooks at https://github.com/huseinzol05/malaya-speech/tree/master/data/noisy-join-mixed-asr
Figure like below,
<img src="/static-proxy?url=https%3A%2F%2Fcdn-lfs.huggingface.co%2Frepos%2Ff9%2F09%2Ff9096f10880558cdf11dc767d127251c9f4d859cebd9b300ab3e98e2117a2157%2F0f9967a068e8681221a92d2e6d4cd8b31f2f50d399b262882f2f41827c196d03%3Fresponse-content-disposition%3Dattachment%253B%2Bfilename*%253DUTF-8%2527%2527augment-noisy-join-mixed-asr.png%253B%2Bfilename%253D%2522augment-noisy-join-mixed-asr.png%2522%253B%26amp%3Bresponse-content-type%3Dimage%252Fpng%26amp%3BExpires%3D1679119880%26amp%3BPolicy%3DeyJTdGF0ZW1lbnQiOlt7IlJlc291cmNlIjoiaHR0cHM6Ly9jZG4tbGZzLmh1Z2dpbmdmYWNlLmNvL3JlcG9zL2Y5LzA5L2Y5MDk2ZjEwODgwNTU4Y2RmMTFkYzc2N2QxMjcyNTFjOWY0ZDg1OWNlYmQ5YjMwMGFiM2U5OGUyMTE3YTIxNTcvMGY5OTY3YTA2OGU4NjgxMjIxYTkyZDJlNmQ0Y2Q4YjMxZjJmNTBkMzk5YjI2Mjg4MmYyZjQxODI3YzE5NmQwMz9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSomcmVzcG9uc2UtY29udGVudC10eXBlPSoiLCJDb25kaXRpb24iOnsiRGF0ZUxlc3NUaGFuIjp7IkFXUzpFcG9jaFRpbWUiOjE2NzkxMTk4ODB9fX1dfQ__%26amp%3BSignature%3DQXze412o9at4ktGZfGpHz4sKL4T9-AFZ7-NVoppv0soj3wbykyyZxN967OxdQjoevVxBtPdsyFt7-vGXCzO-HLyMc5WNtbCCdTqvoEA4UpmeBR7ODYheULvQpFmj3JLrG8dIN36VYnYeMc8oV4Zv%257EjhY524-hM9udqx7r55RfOeNQr%257E0-51f1Vg3UkSob7bECqMXRUWpLV2gKDv9IptVNJAi-QsCIA8a%257ExqqcjXvHzuUs5rJvHLFnkEn4XZc52BboT2Z44PkNT8XO-w4nEnrM8PnHuHrKsK3YtTdOD8Hsqk0zOe4ohK-AGiz621NZulr9LUlsbJQbatKN2YC5hjgnQ__%26amp%3BKey-Pair-Id%3DKVTP0A1DKRTAX%26quot%3B%3C%2Fspan%3E width="50%">
Generated samples ~601 hours. |