DWA-KD: Dual-Space Weighting and Time-Warped Alignment for Cross-Tokenizer Knowledge Distillation
arXiv:2602.21669v1 Announce Type: new Abstract: Knowledge Distillation (KD) has emerged as a crucial technique for compressing Large Language Models (LLMs). Although existing cross-tokenizer KD methods …
Duc Trung Vu, Pham Khanh Chi, Dat Phi Van, Linh Ngo Van, Sang Dinh, Trung Le
조회수 34회