1 repository
Quantization techniques where the spacing between quantization points is optimized via learning.
Distinct from Model Quantization: Specific to non-uniform distributions optimized through back-propagation, distinct from general quantization.
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Non-Uniform Quantization. Refine with filters or upvote what's useful.
PocketFlow is an integrated toolkit for deep learning model compression, distributed training, and mobile format optimization. It provides a system for reducing the size and complexity of neural networks to improve inference efficiency, featuring a dedicated engine for knowledge distillation and a mobile model optimizer. The framework differentiates itself through an automated hyperparameter tuning system that uses reinforcement learning and statistical models to determine optimal compression ratios and layer-wise bit allocation. It also includes a distributed training system that utilizes mu
Optimizes the distribution of quantization points through back-propagation to approximate full-precision network behavior.