@lostprototype Thanks. The team did a great job, and yes, we really do listen. Sometimes, it just takes a bit to work out the right solution.
I will also say we have Full Line Code Completion coming, which provides a local model if you still want some LLM stuff, all at no extra cost to users.
@khalidabuhakmeh - I look forward to it! That's a much more democratized and sensible direction for things to go IMO. It also better connects my annual investment to value, vs. paying for an upsell.
Would that - now or in the future - be able to leverage local GPUs and NPUs for acceleration?
@lostprototype Currently, the FLCC uses a fine-tuned LLaMa model for each language, so there is a baseline for getting a decent experience. I also assume it will work faster when a GPU or NPU is available.
@lostprototype You can download and install FLCC in Rider but the C# model is still being fine-tuned, so results may vary. That’s why we said “coming soon”, we’re not ready to say that it’s ready.
Add comment