anzz1

I'm certainly not versed on how the audio IO / devices actually works but I suspect there must be some input buffer which is then played by the audio devices?...

I'm sorry that on my part I failed to adequately explain my point, and introducing my use case just threw the main point off. As the problem isn't audio monitoring...

> Assuming nobody cares about windows it would be possible to allocate the giant buffer required for the model with [shm_open](https://developer.apple.com/library/archive/documentation/System/Conceptual/ManPages_iPhoneOS/man2/shm_open.2.html) to retain the loaded model in memory between executions....

The problem with WSL2 is not about simplicity, it's about performance. While great improvements in speed have been made recently, a virtualization layer simply can never be as efficient as...

@tarruda I checked out the discussion and I agree that the replacement of STL functions with their standard C runtime counterparts is a good idea, especially the part of using...

This a great idea. As explained in my comment here https://github.com/ggerganov/llama.cpp/issues/23#issuecomment-1477080912 I think the right way going forward would be to separate the concepts of the immutable model and its'...

There is no need to use a separate library to replace C++ fstreams with C file streams, as C file streams work perfectly in Windows too. You are conflating two...

Why even the need of spawning multiple processes instead of multiple threads? Threads are already built-in, unlike sockets or mmap. The only truly portable thing is standard I/O, which can...

@tarruda Please check this comment https://github.com/ggerganov/llama.cpp/issues/23#issuecomment-1477080912 If the TCP server was refactored as a module and use a generic C style of sharing the preloaded model inside the memory of...

Half the fun in AI though is not completely understanding why the results are what they are. I'm only (half) joking though, this will obviously be a good thing. Pitting...