Diagnosing and Enhancing VAE Models

Authors: Bin Dai, David Wipf

ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments from Sections 5 and 6 empirically corroborate motivational theory and reveal that the proposed two-stage procedure can generate high-quality samples...
Researcher Affiliation Collaboration Bin Dai Institute for Advanced Study Tsinghua University Beijing, China daib13@mails.tsinghua.edu.cn David Wipf Microsoft Research Beijing, China davidwipf@gmail.com
Pseudocode No No pseudocode or algorithm blocks were found. The two-stage method is described in narrative text.
Open Source Code Yes The code for our model is available at https://github.com/daib13/Two Stage VAE.
Open Datasets Yes Testing is conducted across four significantly different datasets: MNIST (Le Cun et al., 1998), Fashion MNIST (Xiao et al., 2017), CIFAR-10 (Krizhevsky & Hinton, 2009) and Celeb A (Liu et al., 2015).
Dataset Splits No No explicit train/validation/test dataset splits (e.g., percentages or sample counts) are provided.
Hardware Specification No No specific hardware details (e.g., GPU/CPU models, memory) used for running experiments are provided.
Software Dependencies No All reported FID scores for VAE and GAN models were computed using Tensor Flow (https:// github.com/bioinf-jku/TTUR). No version number specified for TensorFlow. No other software dependencies with version numbers are listed.
Experiment Setup No The paper states 'No effort was made to tune VAE training hyperparameters (e.g., learning rates, etc.); rather a single generic setting was first agnostically selected and then applied to all VAE-like models', but does not provide specific values for these hyperparameters or detailed network architectures used in their experiments.