local MoE model inference

How to Run LFM2.5-24B-A2B Locally on Mac or GPU

Run LiquidAI's LFM2.5-24B-A2B locally via GGUF, MLX, or vLLM. Setup steps, hardware needs, and real throughput numbers from Liquid's own benchmarks.