Whisper by OpenAI is a well-known speech recognition system, but actually it is more than that. It is very powerful but still small enough that it can be used on common hardware. I will provide some background on it's architecture and training before moving on to some experiments, both with the out-of-the box models as well as with the Swedish KB-Whisper model on Swedish speech data.

---

Speakers for Whisper - an introduction:

---

Metadata for Whisper - an introduction

To be recorded: Yes
To be streamed: Yes

---

URLs for Whisper - an introduction

No URLs found.

---

Schedule for Whisper - an introduction

Not scheduled yet