Tag: AI model optimization
Browse our exclusive articles!
PixelPrune: Efficient Visual Token Reduction for VLMs
Discover PixelPrune, a pixel-level adaptive method that speeds up Vision-Language Models by reducing redundant visual tokens without losing accuracy.
Routing-Free Mixture-of-Experts: Scalable AI Model
Discover Routing-Free Mixture-of-Experts, a scalable AI model offering decentralized activation, continuous gradients, and adaptive load balancing.
Efficient Token Pruning for Large Vision Language Models
Discover a novel token pruning method that improves efficiency and performance in large vision language models without extra training.
UniMixer: Unified Scaling Architecture for Recommendation Systems
Discover UniMixer, a unified architecture boosting scaling efficiency and performance in recommendation systems with optimized token mixing.
Efficient Multi-LoRA Generative Vision Models on Edge
Discover QUAD: a unified quantization method enabling multi-LoRA generative vision models with reduced memory and faster performance on edge devices.