Thox-ai/ThoxAir-Accel

docs: temporary pause notice / personal mirror pointer

widen the throughput claim to the measured range (~150-260 tok/s)

correct the throughput claim: ~230 tok/s sustained (~21x), not 1600 (146x)

size llama.cpp threads from the cgroup quota, not os.cpu_count()

install llama-cpp-python prebuilt from conda-forge instead of compiling it

parallelise + narrow the llama.cpp compile (Docker build also hits Job timeout)

convert to sdk:docker -- build llama-cpp-python against this image's glibc

bind GRADIO_SERVER_PORT (7860), not PORT (7861, already bound)

serve the ASGI app with uvicorn -- defining `app` never served it (exit 0)

pin python_version 3.11 for llama-cpp-python cp311 wheels

llama-cpp-python 0.3.19 from the CPU wheel index (0.3.16 had no wheel -> source build timeout)

Upload folder using huggingface_hub

initial commit