WebGPU back end in llama.cpp/ggml

(twitter.com) by tosh | May 22, 2026 | 0 comments on HN