view article Article Making Knowledge Distillation Cheap Enough to Run at Scale MultiverseComputingCAI • about 8 hours ago • 16
Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss Paper • 2608.03796 • Published 7 days ago • 9
Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss Paper • 2608.03796 • Published 7 days ago • 9