{"version":"1.0","type":"video","provider_name":"Bisque","provider_url":"https://bisque.today","title":"llama.cpp v1.0.0 runs MoE models too big for your GPU up to 2× faster","author_name":"@llama-cpp","author_url":"https://bisque.today/p/llama-cpp","html":"<iframe src=\"https://bisque.today/p/llama-cpp/llama-cpp-1-0-0?embed=1\" width=\"800\" height=\"450\" frameborder=\"0\" allow=\"autoplay; fullscreen\" allowfullscreen title=\"llama.cpp v1.0.0 runs MoE models too big for your GPU up to 2× faster\"></iframe>","width":800,"height":450,"thumbnail_url":"https://bisque.today/poster/MexVBeCQUbVVHbCcZgm0IfYg6MJ2/llama-cpp-1-0-0-40d6ac/og-slide.png?v=1794575904337","thumbnail_width":1200,"thumbnail_height":630}