Duration of audio sequence ingested by Whisper

Hi colleagues,

I have an issue when using Whisper. It transcribes only around 30 seconds of audio. Is it a known limitation? How can I ask for the transcription of longer audio files?

Thanks

Best regards

Jerome

Can you please share how you deployed your version?

We published a blog post on how to deploy it, which includes examples with longer audio transcription: Managed Transcription with OpenAI Whisper and Hugging Face Inference Endpoints

Hi Phil,

thanks for your reply. I have just used the inference API (not a specific endpoint deployed for my private usage). Maybe it is the reason. Need I deploy my own endpoint instead?

Thanks

Best regards