The "SPQRAlive" tag likely refers to a specific version or variant in a production pipeline (potentially version 18) optimized for "live" or real-time inference environments. These variants often include:
: It uses a Hessian-based regularizer to identify which weights are most sensitive to quantization. SPQR.SPQRAlive.18.var
: It enables models like LLaMA-65B to fit on a single 24GB or 32GB GPU while maintaining performance. The "SPQRAlive" tag likely refers to a specific