Sleeping PyTorch Weights-only-Quantization (WoQ) 📉 Inference scripts for pytorch weights-only-quantization