Pinned Loading

1. muse-glimmer-dgx-spark muse-glimmer-dgx-spark Public

   Measured Muse Glimmer 30B recipe for NVIDIA DGX Spark, with llama.cpp, DFlash parity checks, vision, tools, and GB10 benchmarks.

   Python 1

2. deepseek-v4-qwen-mm deepseek-v4-qwen-mm Public

   DeepSeek V4 Flash thinks. Qwen-MM sees. Measured stdio MCP vision composition.

   Python

3. gb10-dissection gb10-dissection Public

   Empirical NVIDIA GB10 and SM121 microarchitecture characterization

   Python

4. gx10-nemotron-kit gx10-nemotron-kit Public

   Python

5. nemotron-puzzle-mlx nemotron-puzzle-mlx Public

   Run and benchmark Nemotron Puzzle 75B on 64 GB Apple Silicon with mixed-precision MLX quantization

   Python

6. qwen36-cross-platform qwen36-cross-platform Public

   Reproducible Qwen3.6-35B-A3B benchmark across NVIDIA GB10 and Apple M2 Max, with 4K to 256K context evidence.

   Python