Popular repositories Loading
-
qwen38-flash-next-dgx-spark
qwen38-flash-next-dgx-spark PublicALPHA — Qwen3.8-Flash-Next (176B MoE, vision, 262K ctx) on a single NVIDIA DGX Spark / GB10 with llama.cpp. Pinned build on unmerged PRs; works-on-our-machine, not production-vetted.
Shell 2
-
Archon
Archon PublicForked from coleam00/Archon
Beta release of Archon OS - the knowledge and task management backbone for AI coding assistants.
Python
-
qwen38-flash-next-dgx-spark-1
qwen38-flash-next-dgx-spark-1 PublicForked from sxuff/qwen38-flash-next-dgx-spark
Verified Qwen3.8 Flash-Next GGUF recipe for one NVIDIA DGX Spark or ASUS Ascent GX10
Python
-
lumen-tide
lumen-tide PublicReal-time WebGL 2 bioluminescent sea you can conduct: fluid sim plus generative score, in one self-contained HTML file.
HTML
-
the-long-way-here
the-long-way-here PublicA four and a half minute journey through the 13.8 billion year history of the atoms in your body. One HTML file: WebGL2 and Web Audio, no libraries.
HTML
If the problem persists, check the GitHub status page or contact support.