Multiverse Computing

Latest AI news, models and releases from Multiverse Computing. ['CompactifAI']

Models 🇺🇸

Multiverse Computing: CompactifAI nearly doubles Llama 3.3 performance on Intel Xeon 6

Multiverse Computing's CompactifAI technology nearly doubles the performance of Meta's Llama 3.3 model on Intel Xeon 6 processors, achieving significant speedups without sacrificing accuracy. The solution uses quantum-inspired techniques to optimize AI inference.

Multiverse Computing's AI Model Compression Speeds Up Inference on Standard Hardware

Multiverse Computing has developed a method to compress AI models, enabling them to run faster on standard hardware without specialized accelerators. The technique, which uses singular value decomposition, reportedly achieves up to 10x speedup on CPUs.

Multiverse Computing

Meta AI (GNews)24.07 · 00:03