I uploaded a promt to whisper ui (that is 12 minutes), and it returns “pending”. After I touched transcribe audio, it returns “processing” for ± 10 seconds and then “pending” again. And the site doesn’t work anymore
The base whisper models have a maximum file size of 25 MB that can be processed per request.
From the introductory docs:
By default, the Whisper API only supports files that are less than 25 MB. If you have an audio file that is longer than that, you will need to break it up into chunks of 25 MB’s or less or used a compressed audio format.
You can check your file size to confirm if this is the case here. The documentation also provides some guidelines how to properly chunk long audio files.
When doing this you may want to use the actual whisper prompt to help the model understand the context.
If the file size is actually the culprit, we can still issue a bug report that there should be a proper error message.
https://platform.openai.com/docs/guides/speech-to-text/longer-inputs
When I add files that are more than 25 MB, it is just not allowed to attach a file, so before adding file I checked it for size (the size of current was 20 MB)
Could it be the language problem? Or whisper can recognize any language?
Where exactly / what website are you uploading your files to?
I’m did it to the https:// whisperui . com
឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵឵
That’s not an OpenAI product.
Whisper API can only be consumed using code or the microphone button on Playground in Completions mode
4 Likes
In other words: the website you are using is not a OpenAI website but they are offering OpenAI’s open-sourced whisper model. In this case to upload audio files for processing.
The problem is on their end and not with the whisper model itself.
3 Likes
oh, I thought this is official openai website. Sorry for disturbing