gittensor-ai-lab/sparkinfer

Fastest MoE/LLM inference runtime for consumer and edge Blackwell GPUs. SN74 on Gittensor.

语言：C++ · GitHub 星标：91 · Fork：83

话题：agentic-ai · bittensor · blackwell · cuda · edge-ai · gittensor · gpu · gpu-programming

在 GitHub 上查看原仓库