Lossless Acceleration of LLM via Adaptive N-Gram Parallel Decoding 2 years agoI don't think this will be better than weaker transformer. 0ThreadHN
Lossless Acceleration of LLM via Adaptive N-Gram Parallel Decoding 2 years agohttps://github.com/ggerganov/llama.cpp/tree/master/examples/... 0ThreadHN