Add MLX 4-bit quantized model for iOS on-device inference

f8f9acd

verified

AlonBBar commited on 22 days ago