Audio Speech to Text with openAI whisper model Icelandic to English

[LazyPooky]

The results with AI are exceptional, very accurate and super fast. This topic will discuss the possibilities of AI for the LazyTown Community.

Speech to text

First I would like to mention the transcription of speech to text. I have quite a few videos/audio with the Icelandic language and a lot of them I don't know what is being said. I had an idea about some of them, but it turned out not to be quite right. I use OpenAI with Whisper models. Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. There is now a large-v2 model which does accurate and incredible fast work (on GPU-NVIDIA). The output is simple TXT, and it also creates automatically subtitle files: SRT, VTT .



More here:
https://github.com/openai/whisper
https://huggingface.co/openai/whisper-large

[LazyPooky]

Here are quite a number of whisper models you can choose from, and not only speech-to-text.
https://huggingface.co/models?sort=t...search=whisper

Hugginface is one of the largest AI communities that shares new models and datasets.
https://huggingface.co

[boredjedi]

I'm going to start delving into AI. I been looking into the AI sites.
To the chagrin of Chuft no doubt. I know he's not partial to the AI.
But there's not stopping it now unfortunately. It's either adapt or
be left behind.

[LazyPooky]

I'm going to start delving into AI. I been looking into the AI sites.
To the chagrin of Chuft no doubt. I know he's not partial to the AI.
But there's not stopping it now unfortunately. It's either adapt or
be left behind.

That's great. I try to let AI work locally on Win10, and find software mostly on Github. You don't have to be a nerd or have linux installed. Just read the documentation and follow it. What can be seen online can work local.