Bartłomiej Różański

I found your library and it looks it has a great support for rendering orthogonal edge connections and nodes from dot files however I couldn’t find in Rust ecosystem any...

``` imager_1 | Request: GET /upload/99fdda67-36ba-4b00-af8f-0d612fa1ce97/uploads/719c841d/9d3e/44a8/b014/db007c6b2bd9/original/43351943-aa07b74e-9244-11e8-8e6e-de4f6f13a30f.png.png?thumbnail=190x190&background=white&flatten=true imager_1 | ** (exit) an exception was raised: imager_1 | ** (CaseClauseError) no case clause matching: {:error, :enotconn} imager_1 | (imager) lib/imager_web/controllers/image.ex:34: anonymous fn/2...

Can it highlight reading text?

1

Like in karaoke, would it be somehow doable to highlight which token is spoken?

Would it be possible to combine words with timestamps and perhaps return optionally dict with audio tensor and transcription mapping?

Map architecture in config.json and tokenizer.json files on HuggingFace

1

I've experimented with the Bark and your model and I've found your model simpler to follow and lighter than the Bark model, I'd like to port it in the Elixir...

if dataset is closed https://github.com/suno-ai/bark/issues/2 could you perhaps share the data format for training to make audio model conditioned on the transcript? dataset structure with few examples would be really...