Liquid AI adds small helper models to speed up LFM2.5 output up to 3x

Original: Up to 3.2x Faster Inference with LFM2.5-DSpark

Source: huggingface.co ↗