Popular repositories Loading

1. qwen38-mtp qwen38-mtp Public

   One llama.cpp flag unlocks +33-39% decode speed for Qwen3.8-27B on consumer GPUs. The MTP head already ships inside your GGUF. Recipe, paired benchmarks, probe tool.

   Python 350 99

2. octopus-invaders octopus-invaders Public

   A pixel art space shooter built entirely by a 9B AI model on a single RTX 3060. Zero hand-written code.

   JavaScript 106 11

3. local-ai-on-your-gaming-rtx local-ai-on-your-gaming-rtx Public

   run a 27b local ai on your gaming rtx, 3060, 3070, 4060, 4070, 5060, 5070 and up, 8gb or more, linux and windows, one line to install

   Python 105 12

4. hermes-agent hermes-agent Public

   Forked from NousResearch/hermes-agent

   The agent that grows with you

   Python 61 12

5. nopasswd-sudo nopasswd-sudo Public

   Time-boxed passwordless sudo for agents. Grants expire on a systemd timer, validated with visudo before install, audited with sudoreplay.

   Shell 27

6. dgx-spark-ling dgx-spark-ling Public

   Serve the official Ling-3.0-flash INT4 on one NVIDIA DGX Spark, 38.7 tok/s, faster than the community GGUF. Recipe, watchdog, benchmarks.

   JavaScript 26 1