Learning Frequency Domain Approximation for Binary Neural Networks
Authors: Yixing Xu, Kai Han, Chang Xu, Yehui Tang, Chunjing XU, Yunhe Wang
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The experiments on several benchmark datasets and neural architectures illustrate that the binary network learned using our method achieves the state-of-the-art accuracy. In this section, we evaluate the proposed method on several benchmark datasets such as CIFAR-10 [23] and Image Net [7] on NVIDIA-V100 GPUs to show the superiority of FDA-BNN. |
| Researcher Affiliation | Collaboration | 1Huawei Noah s Ark Lab 2The University of Sydney 3Peking University {yixing.xu, kai.han, tangyehui, xuchunjing, yunhe.wang}@huawei.com c.xu@sydney.edu.au |
| Pseudocode | Yes | Algorithm 1 Feed-Forward and Back-Propagation Process of a convolutional layer in FDA-BNN. |
| Open Source Code | Yes | Code will be available at https://gitee.com/mindspore/models/tree/master/research/cv/FDA-BNN. |
| Open Datasets | Yes | several benchmark datasets such as CIFAR-10 [23] and Image Net [7] |
| Dataset Splits | Yes | The CIFAR-10 [23] dataset contains 50,000 training images and 10,000 test images from 10 different categories. Image Net dataset contains over 1.2M training images with 224 224 resolutions from 1,000 different categories, and 50k validation images. |
| Hardware Specification | Yes | In this section, we evaluate the proposed method on several benchmark datasets such as CIFAR-10 [23] and Image Net [7] on NVIDIA-V100 GPUs |
| Software Dependencies | No | The paper mentions 'All the models are implemented with Py Torch [37] and Mind Spore [21]' but does not provide specific version numbers for these software dependencies. |
| Experiment Setup | Yes | During the experiments, we train the models for 400 epochs with a batch size of 128 and set the initial learning rate as 0.1. The SGD optimizer is used with momentum set of 0.9 and weight decay of 1e-4. We use the Adam optimizer with momentum of 0.9 and set the weight decay to 0. The learning rate starts from 1e-3. |