Custom speech model: data upload error

Question

I am trying to upload data in 'audio + transcript' format.
Getting following error everytime:

Acoustic data import failed: Zero transcriptions could be parsed from the given input.
Error: invalid input line format

Tried copying my human labled text to sample text file. That was successful.
But it fails when uploading my audio files.
Checked the audio file format using SoX. It is correct as per the documentation.

Answer 1

Hi, Can you try using the sample audio+transcript provided in the repo to see if it works. Also, ensure that your data follows the input format and requirements. For example, when uploading training and testing data, the .zip file size cannot exceed 2 GB. Each training file cannot exceed 60 seconds otherwise it will error out. Please review the following document for more details. Hope this helps.

---

--- *Kindly Accept Answer if the information helps. Thanks.*