Exact (token-identical) Llama 3.2 1B Instruct inference on x86 CPUs, and the EAGLE-3 draft head behind it. Code: github.com/ViezeVingertjes/turbo-llm
-
ViezeVingertjes/Llama-3.2-1B-Instruct-EAGLE3
Text Generation • 0.1B • Updated • 571 -
meta-llama/Llama-3.2-1B-Instruct
Text Generation • 1B • Updated • 1.31M • 2k -
EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
Paper • 2503.01840 • Published • 11 -
ViezeVingertjes/Llama-3.2-1B-Instruct-SmolTalk-Self-Distill
Viewer • Updated • 400k • 16