llama.cpp

bd51d63b - Merge pull request #16 from ggml-org/xsn/private_batch_api_pooling_none