MrElmida

I implemented Together AI and LM Studio. For LM Studio, I used the SDK from LM Studio to also implement the speculative decoding feature.

There was a small mistake in the groq.js file, regarding the result output. Instead of getting the content of the output, the whole message object was returned.