juiceb0xc0de/qwen38-27b-cu128-torch280-py311
Updated • 25
SPIN: Simple Plug in Networks
Why fight with Triton kernels and causal-conv1d compilers in your GPaaS?Just grab one of these Docker images and take her for a spin!
-
-
juiceb0xc0de/gemma-4-e2b-it-cu128-torch291-py311
Updated • 26 -
juiceb0xc0de/llama-cpp-cuda130-blackwell100
Preview • Updated • 22
juiceb0xc0de/llama-cpp-cuda128-blackwell120
Preview • Updated • 17Note Pre-compiled llama.cpp binaries for NVIDIA Blackwell GPUs (sm_120 — RTX 50 series, RTX PRO Blackwell). Built with CUDA 12.8 on PyTorch 2.8.0. Verified working via tarball restore.
-
juiceb0xc0de/llama-cpp-cuda12-ada89
Updated • 69 -
juiceb0xc0de/llama-cpp-cuda12-ampere86
Updated • 42 -
juiceb0xc0de/llama-cpp-cuda128-ampere80
Updated • 14