List your company

Speculative Decoding on Specialized Domains Average

114.26Throughput (tokens/s)EvoSpec
100.3656103.9728107.58111.1872Apr 17, 2026

Evaluation results

MethodLinks
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04114.261.07
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04113.111.06
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04112.631.05
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04112.581.05
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04111.151.04
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04110.771.04
FR-SpecStatic Vocabulary Size=32k, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04109.081.02
Standard SDDraft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04106.81
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04104.890.98
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04104.120.97
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04100.90.94