Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks

Authors: Emily L. Denton, Soumith Chintala, arthur szlam, Rob Fergus

NeurIPS 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In a quantitative assessment by human evaluators, our CIFAR10 samples were mistaken for real images around 40% of the time, compared to 10% for samples drawn from a GAN baseline model. We also show samples from models trained on the higher resolution images of the LSUN scene dataset. We evaluate our approach using 3 different methods: (i) computation of log-likelihood on a held out image set; (ii) drawing sample images from the model and (iii) a human subject experiment that compares (a) our samples, (b) those of baseline methods and (c) real images.
Researcher Affiliation Collaboration Emily Denton Dept. of Computer Science Courant Institute New York University Soumith Chintala Arthur Szlam Rob Fergus Facebook AI Research New York
Pseudocode No The paper describes procedures in text and diagrams but does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes Torch training and evaluation code, along with model specification files can be found at http://soumith.ch/eyescream/.
Open Datasets Yes We apply our approach to three datasets: (i) CIFAR10 [17] 32x32 pixel color images of 10 different classes, 100k training samples with tight crops of objects; (ii) STL10 [2] 96x96 pixel color images of 10 different classes, 100k training samples (we use the unlabeled portion of data); and (iii) LSUN [32] 10M images of 10 different natural scene types, downsampled to 64x64 pixels.
Dataset Splits No The paper mentions using a 'validation set' for model selection and log-likelihood comparison, but does not provide specific details on how this set was created (e.g., percentages, sample counts, or splitting methodology).
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions 'Torch training and evaluation code' but does not provide specific version numbers for Torch or any other software dependencies.
Experiment Setup Yes The loss in Eqn. 2 is trained using SGD with an initial learning rate of 0.02, decreased by a factor of (1 + 4 * 10^-4) at each epoch. Momentum starts at 0.5, increasing by 0.0008 at epoch up to a maximum of 0.8.