CP-NAS: Child-Parent Neural Architecture Search for 1-bit CNNs

Authors: Li'an Zhuo, Baochang Zhang, Hanlin Chen, Linlin Yang, Chen Chen, Yanjun Zhu, David Doermann

IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate that the proposed CP-NAS achieves a comparable accuracy with traditional NAS on both the CIFAR and Image Net databases. It achieves the accuracy of 95.27% on CIFAR-10, 64.3% on Image Net with binarized weights and activations, and a 30% faster search than prior arts. 3 Experiments In this section, we compare our CP-NAS with the state-of-the-art NAS methods and 1-bit CNNs methods on two publicly available datasets: CIFAR-10 [Krizhevsky et al., 2014] and ILSVRC12 Image Net [Russakovsky et al., 2015].
Researcher Affiliation Academia Li an Zhuo1 , Baochang Zhang1 , Hanlin Chen1 , Linlin Yang2 , Chen Chen3 , Yanjun Zhu4 and David Doermann4 1School of Automation Science and Electrical Engineering, Beihang University 2University of Bonn 3University of North Carolina at Charlotte 4University at Buffalo {lianzhuo, bczhang, hlchen}@buaa.edu.cn
Pseudocode Yes Algorithm 1 Child-Parent NAS Input: Training data, Validation data Parameter: Searching hyper-graph: G, K = 8, e(o(i,j) k ) = 0 for all edges Output: Optimal structure α 1: while (K > 1) do 2: for t = 1, ..., T epoch do ...
Open Source Code No The paper does not include a specific link to source code or an explicit statement about the release of code for the described methodology.
Open Datasets Yes 3 Experiments In this section, we compare our CP-NAS with the state-of-the-art NAS methods and 1-bit CNNs methods on two publicly available datasets: CIFAR-10 [Krizhevsky et al., 2014] and ILSVRC12 Image Net [Russakovsky et al., 2015].
Dataset Splits Yes During the architecture search, the training set of the dataset is divided into two subsets, one for training the network weights and the other for perfomrance evaluation as a validation set. ... Due to the efficient guidance of CP model, we only use 50% of the training set with CIFAR-10 and Image Net for architecture search and 5% of the training set for evaluation, leading to a faster search.
Hardware Specification Yes In terms of search efficiency, compared with the previous work PC-DARTS [Xu et al., 2019], our CP-NAS is 30% faster (tested on our platform 6 NVIDIA TITAN V GPUs).
Software Dependencies No The paper states 'All the experiments and models are implemented in Py Torch [Paszke et al., 2017]' but does not provide specific version numbers for PyTorch or any other software dependencies.
Experiment Setup Yes We use SGD with momentum to optimize the network weights, with an initial learning rate of 0.025 (annealed down to zero following a cosine schedule), a momentum of 0.9, and a weight decay of 5 × 10−4. When we search for the architecture directly on Image Net, we use the same parameters for searching with CIFAR-10 except that the initial learning rate is set to 0.05 and βP is set to 0.33. A larger network of 10 cells ... is trained on CIFAR-10 for 600 epochs with a batch size of 96 ... We use the SGD optimizer with an initial learning rate of 0.025 ... a momentum of 0.9, a weight decay of 3 × 10−4 and a gradient clipping at 5.