Inference code for EAGLE-4, a speculative decoding harness that speeds up LLM serving on SGLang, vLLM and TokenSpeed.
2296PythonOtherUpdated 10 hours ago
#eagle
Curated GitHub repositories related to eagle, with AI summaries and key stats.
Inference code for EAGLE-4, a speculative decoding harness that speeds up LLM serving on SGLang, vLLM and TokenSpeed.