Differentiable Model Compression via Pseudo Quantization Noise

Open paper
Benchmark Model Rank Results
language-modelling-on-wikitext-103DIFFQ (λ=1, g=16)#33Test perplexity: 18.0
image-classification-on-imagenetDIFFQ (λ=1e−2)#617Top 1 Accuracy: 82.0

Rank counts only results with a code link.