Cross Attention Network for Few-shot Classification
Authors: Ruibing Hou, Hong Chang, Bingpeng MA, Shiguang Shan, Xilin Chen
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on two benchmarks show our method is a simple, effective and computationally efficient framework and outperforms the state-of-the-arts. |
| Researcher Affiliation | Academia | Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS), Institute of Computing Technology, CAS, China 2University of Chinese Academy of Sciences, China 3CAS Center for Excellence in Brain Science and Intelligence Technology, China ruibing.hou@vipl.ict.ac.cn, {changhong, sgshan,xlchen}@ict.ac.cn, bpma@ucas.ac.cn |
| Pseudocode | No | The paper describes the proposed methods in narrative text and uses diagrams, but does not include formal pseudocode or algorithm blocks. |
| Open Source Code | Yes | The code and models are available on https://github.com/blue-blue272/fewshot-CAN |
| Open Datasets | Yes | We use mini Image Net [40] which is a subset of ILSVRC-12 [13]. It contains 100 classes with 600 images per class. We use the standard split following [31, 37, 26, 15, 33]: 64 classes for training, 16 for validation and 20 for testing. We also use tiered Image Net dataset [32], a much larger subset of ILSVRC-12 [13]. |
| Dataset Splits | Yes | We use the standard split following [31, 37, 26, 15, 33]: 64 classes for training, 16 for validation and 20 for testing. We also use tiered Image Net dataset [32]... These are divided into 20 categories (351 classes) for training, 6 categories (97 classes) for validation, and 8 categories (160 classes) for testing, as in [32, 7, 35, 37]. |
| Hardware Specification | Yes | Pytorch [28] is used to implement all our experiments on one NVIDIA 1080Ti GPU. |
| Software Dependencies | No | The paper mentions 'Pytorch [28]' as the implementation framework, but does not provide specific version numbers for Pytorch or any other software dependencies. |
| Experiment Setup | Yes | The input images size is 84 84. During training, we adopt horizontal flip, random crop and random erasing [49] as data augmentation. SGD is used as the optimizer. Each mini-batch contains 8 episodes. The model is trained for 80 epochs, with each epoch consisting of 1, 200 episodes. For mini Image Net, the initial learning rate is 0.1 and decreased to 0.006 and 0.0012 at 60 and 70 epochs, respectively. For tiered Image Net, the initial learning rate is set to 0.1 with a decay factor 0.1 at every 20 epochs. The temperature hyperparameter (τ in Eq. 3) is set to 0.025, the reduction ratio in the meta-learner is set to 6, and the weight hyperparameter (λ) in the overall loss function is set to 0.5. For the transductive algorithm, the selected number of query samples in the first iteration (t) is set to 35, and the number of iterations and enlarging factor of candidate set are both set to 2. |