Pinned Loading

1. ntransformer ntransformer Public

   High-efficiency LLM inference engine in C++/CUDA. Run Llama 70B on RTX 3090.

   C++ 462 20

2. gpu-nvme-direct gpu-nvme-direct Public

   GPU-initiated NVMe I/O via PCIe BAR MMIO — CUDA kernels directly issue NVMe commands, eliminating CPU from the storage data path

   Cuda 22 1

3. brandon-tiny brandon-tiny Public

   Ultra-small instruction-following language models (10M-110M params) that run on a PlayStation 2. BLiMP 73.3%, HellaSwag 32.4% at just 10.7M parameters.

   Python 2

4. ps2-llm ps2-llm Public

   Running a large language model on a PlayStation 2

   C 42 1

5. osito-k osito-k Public

   Bare-metal preemptive kernel for ESP8266 — 3D wireframe Elite demo, zForth interpreter, OsitoFS, all in 24KB of IRAM

   C++ 2

6. ter ter Public

   Ternary computing transformer inference

   Cuda