Quantization-Aware Healing:a compressed 4-bit model that outperf full-precision

2 points | by grigio 4 hours ago