Closed HakaishinShwet closed 5 months ago
Share the file
And the way you run the decoding
i have attached both whisper model result in first two images and after that you can see how vosk model process the music file and generate the file in which nothing useful is there you can in last image, plus in cli too it is not detecting and showing perfect lines of music at all and i tried with different ones too and got same result @nshmyrev
if you wanna test yourself you can download linkin park music from net and test for yourself or any other file you can test and tell if i am doing something wrong because i got this command from documentation and guides
Ok, that kind of task is beyond our current capabilities for now. For the future we might look into it
@nshmyrev ok, hope it atleast reach to open ai whisper model level because their small models are also working very well even on my potato laptop haha so i was expecting not that level but still i was expecting some decent lines generation but it completely failed so i thought like am i doing something wrong or the project development is stopped or is not capable enough yet
getting random text generation for audio file and music file. i mean it is not even generating and getting stuck in between result generation too.Compared to openai whisper it is performing in worst possible way