Turbo-fieldfare: Open-source engine running Gemma 4 26B in 2 GB RAM on Apple Silicon
英文标题：Turbo-fieldfare: Open-source engine running Gemma 4 26B in 2 GB RAM on Apple Silicon
Its a custom Swift/Metal inference engine that runs Gemma 4 26B-A4B-IT on M-series Macs with very low RAM. It uses ~2GB instead of ~14 GB. The result is reportedly 5–6 tok/s on an 8 GB M2 MacBook Air and 31–35 tok/s on an M5 MacBook Pro. It also includes an OpenAI-compatible local server with
打开原文相关 4 · 机会 3
中文
Its a custom Swift/Metal inference engine that runs Gemma 4 26B-A4B-IT on M-series Macs with very low RAM. It uses ~2GB instead of ~14 GB. The result is reportedly 5–6 tok/s on an 8 GB M2 MacBook Air and 31–35 tok/s on an M5 MacBook Pro. It also includes an OpenAI-compatible local server with
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
Its a custom Swift/Metal inference engine that runs Gemma 4 26B-A4B-IT on M-series Macs with very low RAM. It uses ~2GB instead of ~14 GB. The result is reportedly 5–6 tok/s on an 8 GB M2 MacBook Air and 31–35 tok/s on an M5 MacBook Pro. It also includes an OpenAI-compatible local server with