Steward Garcia

That would require retraining the model, or one could possibly convert an existing one. I didn't quite understand that part of the paper.

Waiting for some update

This should be tested in all platforms

Hello, cleaning emb_inp you don't resetting the llama context, In `kv_cache`, remnants of the previous context may remain, so resetting the context in this way may not yield the best...

> What is `interactive: true` supposed to do? That option allows for interactions with the model; it stops generating text when it encounters a `stop word`, whereas when it is...

@SlyEcho > * [x] La semilla RNG debe ser un parámetro JSON, no un argumento CLI del servidor, si no está configurado, la misma lógica. I had considered that before,...

> Intente que la IA envíe caracteres de varios bytes como un emoji acaba en un error con el estado 500, no hay ningún mensaje de error ni nada impreso...

Hello, if you provide me with more information, I could help you. If you could provide a screenshot of the error, it would be even better, especially if it's a...

Reviewing your comment again, I suggest you do not use the latest version of ggml (Anyway, there is no significant improvement in any aspect) as it introduces many changes that...

> this personal AI study project is an Android turn-key project, you can reproduce this issue very easily on Xiaomi 14 or any other mainstream Android phone. I will try...