SOL Attention plus SageAttention speeds Apple Silicon inference 2.5x

A developer combined SOL Attention with a SageAttention-style INT8 path in Vpipe, achieving up to 2.5x faster H3 inference on Apple Silicon.

2026-09-10 ~ 2026-09-10 · 2 related posts

• SOL Attention + SageAttention Speeds Up H3 Video Gen Up to 2.5x on Apple Silicon — TgoAI · 2026-09-10
• SOL-H3 with SageAttention delivers up to 2.5x faster H3 inference on Apple Silicon — TgoAI · 2026-09-10