10-11-2026, 06:56 AM
Interesting, I hadn't tried that. The pattern around continuous batching has been pretty similar for me.
|
speculative decoding in llama.cpp with SGLang
|
|
« Next Oldest | Next Newest »
|