Pinned Loading
-
strata-gfx906
strata-gfx906 PublicStrata 0.2.3 (engine 0.1.48) for AMD gfx906 (MI50/MI60/Radeon VII): rocBLAS solution table, Qwen3.6/Ornith engine restored, DFlash2 build fix, web configurator, portable (no AVX-VNNI needed). Docke…
-
llama-mi50-rocm7.2
llama-mi50-rocm7.2 Publicllama.cpp b12128 (0.9.0-dev) for AMD gfx906 (MI50/MI60/Radeon VII): 14 CPU backend variants, patched gfx906 rocBLAS, DeepSeek-V4-Flash (deepseek4/deepseek41) support. Docker: xxdoman/llama-mi50-roc…
Shell
-
maya-gfx906
maya-gfx906 PublicMaya dashboard on a gfx906 (MI50/MI60/Radeon VII) build of the Strata engine: DeepSeek-V4-Flash-0731 at 13.3 tok/s on one MI50. Docker: xxdoman/maya-gfx906
Dockerfile
-
ollama-mi50-rocm7.2-optimized-gfx906
ollama-mi50-rocm7.2-optimized-gfx906 PublicOptimized Ollama (v0.40.14) Docker image for AMD Instinct MI50 (gfx906), ROCm 7.2 - 13 CPU backend variants, LFM3 support
-
ollama-amd-mi50
ollama-amd-mi50 PublicOllama 0.12.3 + ROCm 7.1 for AMD MI50 (gfx906) Docker image
-
If the problem persists, check the GitHub status page or contact support.
