jayleaton/deepseek-v41-tensorfold-spark

DeepSeek-V4.1-Flash on 2x NVIDIA DGX Spark with the TensorFold engine: 1.8-1.9x code, 2.5x multi-stream, 1.7-1.95x prefill vs the vLLM kit, exact speculative decoding. Work in progress.

语言：Zig · GitHub 星标：116 · Fork：18

在 GitHub 上查看原仓库