Constructing Fast Network through Deconstruction of Convolution

Authors: Yunho Jeon, Junmo Kim

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To demonstrate the performance of our proposed method, we conducted several experiments with classification benchmark datasets.
Researcher Affiliation Academia Yunho Jeon School of Electrical Engineering, KAIST jyh2986@kaist.ac.kr Junmo Kim School of Electrical Engineering, KAIST junmo.kim@kaist.ac.kr
Pseudocode No The paper describes methods through equations and text but does not include any clearly labeled 'Pseudocode' or 'Algorithm' blocks.
Open Source Code Yes Code is available at https://github.com/ jyh2986/Active-Shift.
Open Datasets Yes We conducted experiments to verify the basic performance of ASL with the CIFAR-10/100 dataset [14] that contains 50k training and 10k test 32 32 images. To prove the generality of the proposed method, we conducted experiments with an Image Net 2012 classification task.
Dataset Splits No The paper states using '50k training and 10k test' images for CIFAR-10/100 but does not explicitly provide details for a validation split.
Hardware Specification Yes Time is measured using an Intel i7-5930K CPU with a single thread and averaged over 100 repetitions. Measured by Caffe [12] using an Intel i7-5930K CPU with a single thread and GTX Titan X (Maxwell).
Software Dependencies No The paper mentions using 'Caffe [12]' but does not specify a version number for the software dependency.
Experiment Setup Yes For ASL, the shift parameters are randomly initialized with uniform distribution between -1 and 1. We used a normalized gradient following ACU[11] with an initial learning rate of 1e-2. Input images are normalized for all experiments.