llama.cpp prebuilt b9739-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9739, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9726-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9726, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9682-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9682, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9670-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9670, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9654-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9654, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9651-mix-2d6bd50

Automated Unsloth llama.cpp CUDA + ROCm + Vulkan + macOS + CPU prebuild for upstream b9651, merged with:

• DiffusionGemma (#24423, commit 49fc372)
• Preliminary MiniMax-M3 support (#24523, commit 2846a38)

llama.cpp prebuilt b9625-mix-2d6bd50

Automated Unsloth llama.cpp CUDA (Linux + Windows) + ROCm + macOS + CPU (Linux + Windows) + Vulkan (Linux + Windows) prebuild for upstream b9625, merged with:

• ggml-org#24423 @ 49fc372
• ggml-org#24523 @ 2846a38

llama.cpp prebuilt b9616-mix-5efc536

Automated Unsloth llama.cpp CUDA (Linux + Windows) + ROCm + macOS + CPU (Linux + Windows) + Vulkan (Linux + Windows) prebuild for upstream b9616 + PR ggml-org#24423 @ 1153c4a, PR ggml-org#24523 @ 2846a38, PR ggml-org#24260 @ d932047 (10196c0).

llama.cpp prebuilt b9590-mix-f4d21e7

Automated Unsloth llama.cpp CUDA (Linux + Windows) + ROCm + macOS prebuild for upstream b9590 + PR ggml-org#24423 @ c84e85a (feb4f53).

llama.cpp prebuilt b9587-mix-f4d21e7

Automated Unsloth llama.cpp CUDA (Linux + Windows) + ROCm + macOS prebuild for upstream b9587 + PR ggml-org#24423 @ c84e85a (ggml-org/llama.cpp@ec32531).