4 comments

[ 0.29 ms ] story [ 23.5 ms ] thread
We are proud to launch our game-changing llama-cpp-wasm engine, a breakthrough in-browser LLM engine that executes on the edge using general purpose computing based on llama.cpp.

Desktop browser version is OK. Mobile browser version is WIP.

Does it use webgpu?
Nope, it is just Wasm/JS. Please test it on desktop Firefox multi-threading version since it turns out to be the fastest one.
How does the performance of the desktop versions of Chrome, Safari, and Firefox compare?