AFM Core Advanced on-device model running on A19 Pro is a sparse model.
It's 20B parameters.
It's fully Apple designed. It is an MoE but when it processes the prompt, it only loads the parameters needed and locks them in.
If it's 20B parameters total, but on a specific request it's only 1-4B parameter total. It only loads in 1-4B for inference and decides them at prefill time.
It is fully Apple designed architecture, Google had nothing here.
今回は生成AIの心臓であるAttentionを解説をしています。Attentionと聞くとAttention Is All You Needが超有名ですが、実はこれが初出ではありません。今回はその初出に焦点を当てて解説しました。
第14回 Attention機構の誕生:AIが「どこに注目すべきか」を学習する仕組み https://t.co/it34DRkqrj