<ul><li>Vector-quantized networks (VQNs) have shown great performance but suffer from training instability.</li><li>Researchers propose OptVQ, a vector quantization method that integrates optimal transport to improve stability and efficiency of training.</li><li>OptVQ uses the Sinkhorn algorithm to optimize the optimal transport problem.</li><li>Experiments demonstrate that OptVQ achieves 100% codebook utilization and outperforms current state-of-the-art VQNs in image reconstruction quality.</li></ul>

Preventing Local Pitfalls in Vector Quantization via Optimal Transport

Discover more