Adaptive Dual Guidance Knowledge Distillation
Authors: Tong Li, Long Liu, Kang Liu, Xin Wang, Bo Zhou, Hongguang Yang, Kai Lu
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on CIFAR-100, Image Net, and MS-COCO demonstrate the effectiveness of our method. Moreover, our method can be integrated with other KD methods, boosting their performance. |
| Researcher Affiliation | Academia | Tong Li, Long Liu*, Kang Liu, Xin Wang, Bo Zhou, Hongguang Yang, Kai Lu Xi an University of Technology, Xi an, 710048, China EMAIL, EMAIL, Kang EMAIL, EMAIL, EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes the proposed method using mathematical formulations and descriptive text, but it does not include any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code or provide a link to a code repository. |
| Open Datasets | Yes | CIFAR-100 (Krizhevsky, Hinton et al. 2009) is a medium-scale image classification dataset... Image Net (Deng et al. 2009) is one of the most important benchmark datasets for image classification... MS-COCO (Lin et al. 2014) is a fundamental object detection dataset... |
| Dataset Splits | Yes | CIFAR-100 ... consisting of 60,000 images (50,000 training samples and 10,000 testing samples) from 100 categories... Image Net ... with a total of 1.28 million training samples and 50,000 testing samples from 1000 categories. MS-COCO ... with 118k images to train and 5k images to test from 80 categories. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU models, CPU types, or other computer specifications used for running the experiments. |
| Software Dependencies | No | The paper mentions using Detectron2 (Wu et al. 2019) as a baseline report but does not list specific software dependencies with version numbers for the implementation of ADG-KD. |
| Experiment Setup | No | The paper describes the overall framework, loss functions, and adaptive dual guidance approaches. However, it does not explicitly provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or other detailed system-level training settings in the main text. |