Faster Text Generation with Self-Speculative Decoding

T1Covered by 1 source · first reported by Hugging Face Blog at 20 Nov 2024, 00:00 UTC