Vector quantization is common in deep models, yet its hard assignments block gradients and hinder end-to-end training. We propose DiVeQ, which treats quantization as adding an error vector that mimics the quantization distortion, keeping the forward pass hard while letting gradients flow. We also present a space-filling variant (SF-DiVeQ) that assigns input to a curve constructed by the lines connecting codewords, resulting in less quantization error and full codebook usage. Both methods train end-to-end without requiring auxiliary losses or temperature schedules. In VQ-VAE image compression, VQGAN image generation, and DAC speech coding tasks across various data sets, our proposed methods improve reconstruction and sample quality over alternative quantization approaches.
@article{arxiv.2509.26469,
title = {DiVeQ: Differentiable Vector Quantization Using the Reparameterization Trick},
author = {Mohammad Hassan Vali and Tom Bäckström and Arno Solin},
journal= {arXiv preprint arXiv:2509.26469},
year = {2026}
}