#Model Optimization

3 articles with this tag

Nvidia's Ziv Ilan on Faster Diffusion Models
AI Research

Nvidia's Ziv Ilan on Faster Diffusion Models

Nvidia's Ziv Ilan explains how to reduce diffusion model latency using quantization, caching, and distillation, plus the new FastGen library.

2 months ago
DoRA Efficiency Breakthrough
AI Research

DoRA Efficiency Breakthrough

New factored norm and fused kernels unlock DoRA's potential, delivering 1.5-2x speedups and significant VRAM reduction.

5 months ago
Pretraining's Hidden Experts: A New Post-Training Paradigm
AI Research

Pretraining's Hidden Experts: A New Post-Training Paradigm

Large pretrained models are dense with task-experts, enabling simple random sampling and ensembling to rival complex post-training AI optimization methods.

5 months ago