Liquid AI shrinks models to 4-bit with near-full accuracy

Original: LFM2.5 Q4\_0 Checkpoints from Quantization-Aware Distillation

Source: huggingface.co ↗