Topic: embedding

The ggml/llama.cpp project released version b10199, which adds support for input embedding to generate the next token and fixes issues with server_batch(). These updates enhance the model's functionality and performance.