作者位次 #5/10ACL 2026Poster· Natural Language Processing
BTC - LLM : Efficient Sub-1-Bit LLM Quantization via Learnable Transformation and Binary Codebook
Hao Gu · Lujun Li · Hao Wang · Lei Wang · Zheyu Wang · Bei Liu · Jiacheng Liu · Qiyuan Zhu · Sirui Han · Yike Guo
摘要
Binary quantization represents the most extreme form of compression, reducing weights to \pm 1 for maximal memory and computational efficiency. While recent sparsity-aware binarization achieves sub-1-bit compression via weight pruning, it faces critical challenger: performance degradation, mask-m...