Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook
Swift apple-silicongemmagemma4gemma4-26b-a4bgpgpullmllm-inferencelocal-aimacosmetalon-device-aion-device-llmswift
| Name | Last commit message | Last commit date |
|---|---|---|
| .. | ||
| FusedLayerTail.swift 2.5 KB | ||
| FusedPostAttentionSetup.swift 2.9 KB | ||
| FusedQKVEpilogue.swift 4.3 KB | ||
| FusedQKVGEMV.swift 4.4 KB | ||
| LMHeadChainInt4.swift 5.6 KB | ||