3 papers
cs.CV2026
NSVQ: Mitigating Codebook Collapse by Stabilizing Encoder Drift in Vector Quantization
Hao Lu, Yongxin Guo, Onur Koyun +3
Vector quantization is central to modern generative modeling pipelines, but large-codebook VQ models often suffer from codebook collapse. We identify encoder drift as a key driver…
cs.LG2026
PCA-VAE: Differentiable Subspace Quantization without Codebook Collapse
Hao Lu, Onur C. Koyun, Yongxin Guo +3
Vector-quantized autoencoders deliver high-fidelity latents but suffer inherent flaws: the quantizer is non-differentiable, requires straight-through hacks, and is prone to collaps…
cs.CV2026
Beyond Stationarity: Rethinking Codebook Collapse in Vector Quantization
Hao Lu, Onur C. Koyun, Yongxin Guo +3
Vector Quantization (VQ) underpins many modern generative frameworks such as VQ-VAE, VQ-GAN, and latent diffusion models. Yet, it suffers from the persistent problem of codebook co…