llama.cpp

d4e0bad0 - server : (embeddings) using same format for "input" and "content"