Download¶
MBASE SDK v0.1.0 download page
Bundled SDK¶
If you don’t have the llama.cpp on your system, go with the bundled SDK which will install both llama.cpp and MBASE SDK with proper configuration.
x86_64, Apple Metal
x86 CPU, NVIDIA CUDA GPU
Find your NVIDIA GPU: https://developer.nvidia.com/cuda-gpus
NVIDIA Datacenter GPUs Compatible
NVIDIA Quadro and NVIDIA RTX GPUs Compatible
NVIDIA GeForce and Titan GPUs Compatible
NVIDIA Jetson GPUs Compatible
Standalone SDK¶
Only the MBASE SDK. If you have a llama.cpp library already installed on your system, go with the standalone SDK.