llama.cpp VS rust-gpu

Have just done this recently for local chat with pdf feature in https://recurse.chat. (It's a macOS app that has built-in llama.cpp server and local vector database)

Running an embedding server locally is pretty straightforward:

- Get llama.cpp release binary: https://github.com/ggerganov/llama.cpp/releases

Ah, thanks for this! I can't edit my parent comment that you replied to any longer unfortunately.

As I said, I only compared the contributors graphs [0] and checked for overlaps. But those apparently only go back about year and only list at most 100 contributors ranked by number of commits.

[0]: https://github.com/ollama/ollama/graphs/contributors and https://github.com/ggerganov/llama.cpp/graphs/contributors

KodiBot is a desktop app that enables users to run their own AI chat assistants locally and offline on Windows, Mac, and Linux operating systems. KodiBot is a standalone app and does not require an internet connection or additional dependencies to run local chat assistants. It supports both Llama.cpp compatible models and OpenAI API.

There are already some implementations out there which attempt to accomplish this!

Here's an example: https://github.com/silphendio/sliced_llama

A gist pertaining to said example: https://gist.github.com/silphendio/535cd9c1821aa1290aa10d587...

Here's a discussion about integrating this capability with ExLlama: https://github.com/turboderp/exllamav2/pull/275

And same as above but for llama.cpp: https://github.com/ggerganov/llama.cpp/issues/4718#issuecomm...

For those who might not be aware of this, there is also an open source project on GitHub called "Twinny" which is an offline Visual Studio Code plugin equivalent to Copilot: https://github.com/rjmacarthy/twinny

It can be used with a number of local model services. Currently for my setup on a NVIDIA 4090, I'm running both the base and instruct model for deepseek-coder 6.7b using 5_K_M Quantization GGUF files (for performance) through llama.cpp "server" where the base model is for completions and the instruct model for chat interactions.

llama.cpp: https://github.com/ggerganov/llama.cpp/

deepseek-coder 6.7b base GGUF files: https://huggingface.co/TheBloke/deepseek-coder-6.7B-base-GGU...

deepseek-coder 6.7b instruct GGUF files: https://huggingface.co/TheBloke/deepseek-coder-6.7B-instruct...

If I'm reading this correctly, they had to discard Llama 2 answers and only use GPT-3.5 given answers to test the hypothesis.

GPT-3.5 answering questions through the OAI API alone is not an acceptable method of testing problem solving ability across a range of temperatures. OpenAI does some blackbox wizardry on their end.

There are many complex and clever sampling techniques for which temperature is just one (possibly dynamic) component

One example from the llama.cpp codebase is dynamic temperature sampling

https://github.com/ggerganov/llama.cpp/pull/4972/files

Not sure what you mean by whole model state given that there are tens of thousands of possible tokens and the models have billions of parameters in XX,XXX-dimensional space. How many queries across how many sampling methods might you need? Err..how much time? :)

Sounds cool, but this requires yet another language to learn[0]. As someone who only has limited knowledge in this space, could someone tell me how comparable is the compute functionality of rust-gpu[1], where I can just write rust?

[0] https://github.com/Hugobros3/shady#language-syntax

[1] https://github.com/EmbarkStudios/rust-gpu

I don't do anything related to data science, but I feel like doing it in Rust would be nice.

You get operator overloading, so you can have ergonomic matrix operations that are typed also. Processing data on the CPU is fast, and crates like https://github.com/EmbarkStudios/rust-gpu make it very ergonomic to leverage the GPU.

I like this library for creating typed coordinate spaces for graphics programming (https://github.com/servo/euclid), I imagine something similar could be done to create refined types for matrices so you don't do matrix multiplication matrices of invalid sizes

Do you mean rust-gpu?

And https://github.com/EmbarkStudios/rust-gpu/tree/main/examples with the wgpu runner (here it runs the compute shader)

I don't know how major they are considered, but Embark Studios is doing quite a bit of Rust in the open source space, most notably (IMO) rust-gpu and kajiya

The examples in the rust-gpu repository are a good place to start

There's another project that's similar that's being used by an actual game company: https://github.com/EmbarkStudios/rust-gpu

They see specific advantages here that would outweigh that negative. It's not my space (I play games, but know next to nothing about graphics programming), but there's at least one argument in the other direction.

Could this approach work for compute shaders (GPGPU) as well? So far, I think https://github.com/EmbarkStudios/rust-gpu is the state of the art in that area, but it adds a specific Rust compiler backend for generating SPIR-V rather than leaving that up to the driver. That seems more complicated than it needs to be... but maybe it has advantages too? Thoughts?

https://github.com/embarkstudios/rust-gpu Allows you to create shaders (kernals) in Rust.

Also Embark Studios (formers DICE people) is doing a lot of work with Rust, all open source like Rust GPU https://github.com/EmbarkStudios/rust-gpu