Multi-marginal Wasserstein GAN

Authors: Jiezhang Cao, Langyuan Mo, Yifan Zhang, Kui Jia, Chunhua Shen, Mingkui Tan

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on toy and real-world datasets demonstrate the effectiveness of MWGAN.
Researcher Affiliation Academia Jiezhang Cao , Langyuan Mo , Yifan Zhang, Kui Jia, Chunhua Shen, Mingkui Tan South China University of Technology, Peng Cheng Laboratory, The University of Adelaide {secaojiezhang, selymo, sezyifan}@mail.scut.edu.cn {mingkuitan, kuijia}@scut.edu.cn, chunhua.shen@adelaide.edu.au
Pseudocode Yes Algorithm 1 Multi-marginal WGAN.
Open Source Code Yes The source code of our method is available: https://github.com/caojiezhang/MWGAN.
Open Datasets Yes Datasets. We conduct experiments on three datasets. Note that all images are resized as 128 128. ... (ii) Celeb A [33] contains 202,599 face images, where each image has 40 binary attributes. ... (iii) Style painting [51].
Dataset Splits No The paper does not explicitly provide the training, validation, and test splits for the main datasets (Toy, Celeb A, Style painting) used for the GAN model training. It only mentions a 90% training and 10% testing split for an auxiliary classifier trained on Celeb A.
Hardware Specification Yes All experiments are conducted based on Py Torch, with an NVIDIA TITAN X GPU.
Software Dependencies No The paper mentions 'Py Torch' but does not specify its version or any other software dependencies with version numbers.
Experiment Setup Yes We use Adam [29] with β1=0.5 and β2=0.999 and set the learning rate as 0.0001. We train the model 100k iterations with batch size 16. We set α=10, τ=10 and Lf to be the number of target domains in Loss (7).