Papers › Preventing Local Pitfalls in Vector Quantization via Optimal Transport
Preventing Local Pitfalls in Vector Quantization via Optimal Transport
Borui Zhang, Wenzhao Zheng, Jie zhou, Jiwen Lu
Vector-quantized networks (VQNs) have exhibited remarkable performance across various tasks, yet they are prone to training instability, which complicates the training process due to the necessity for techniques such as subtle initialization and model distillation. In this study, we identify the local minima issue as the primary cause of this instability. To address this, we integrate an optimal transport method in place of the nearest neighbor search to achieve a more globally informed assignment. We introduce OptVQ, a novel vector quantization method that employs the Sinkhorn algorithm to optimize the optimal transport problem, thereby enhancing the stability and efficiency of the training process. To mitigate the influence of diverse data distributions on the Sinkhorn algorithm, we implement a straightforward yet effective normalization strategy. Our comprehensive experiments on image reconstruction tasks demonstrate that OptVQ achieves 100% codebook utilization and surpasses current state-of-the-art VQNs in reconstruction quality.
Syntology This paper's arXiv id is not in Syntology's Atlas id list (549,449 ids, as this build read them; the read time is in the build record), so this page links to no Syntology page for it. The list is not the whole graph; Syntology's feed shows papers as they enter the graph.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Image Reconstruction | ImageNet | OptVQ (16x16x8) | FID | 0.91 | #4 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x8) | LPIPS | 0.066 | #4 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x8) | PSNR | 27.57 | #4 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x8) | SSIM | 0.729 | #4 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x4) | FID | 1.00 | #5 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x4) | LPIPS | 0.076 | #5 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x4) | PSNR | 26.59 | #5 of 15 | Archive leaderboard | report |
| Image Reconstruction | ImageNet | OptVQ (16x16x4) | SSIM | 0.717 | #5 of 15 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections