FineFMPL: Fine-grained Feature Mining Prompt Learning for Few-Shot Class Incremental Learning
Authors: Hongbo Sun, Jiahuan Zhou, Xiangteng He, Jinglin Xu, Yuxin Peng
IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on three benchmark datasets demonstrate that our Fine FMPL achieves new state-of-the-art. |
| Researcher Affiliation | Academia | Hongbo Sun1 , Jiahuan Zhou1 , Xiangteng He1 , Jinglin Xu2 and Yuxin Peng1 1Wangxuan Institute of Computer Technology, Peking University 2School of Intelligence Science and Technology, University of Science and Technology Beijing {sunhongbo, jiahuanzhou, hexiangteng}@pku.edu.cn, xujinglinlove@gmail.com, pengyuxin@pku.edu.cn |
| Pseudocode | No | The paper describes the model architecture and processes in prose and with diagrams, but it does not include any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | Yes | The code is available at https://github.com/PKU-ICST-MIPL/Fine FMPL IJCAI2024. |
| Open Datasets | Yes | We conduct extensive comparison experiments and ablation studies on three standard few-shot class incremental learning (FSCIL) benchmark datasets, i.e., CUB-200-2011 [Wah et al., 2011], CIFAR 100 [Krizhevsky et al., 2009], and mini Image Net [Russakovsky et al., 2015]. These are well-known, publicly available datasets. |
| Dataset Splits | Yes | For fair comparisons with state-of-the-art (SOTA) FSCIL methods, the same benchmark datasets and FSCIL setting [Tao et al., 2020] are adopted, as shown in Table 1. Table 1 provides 'Cbase', 'Cinc', '#Inc', 'N-way-K-shot' for each dataset, which defines the base classes, incremental classes, number of incremental sessions, and the few-shot setting (e.g., 10-way-5-shot for CUB-200-2011), implicitly defining the training and evaluation splits across sessions. For instance, CUB-200-2011 has '100 classes as base classes, and the remaining classes are split into 10 sessions, where each session learns 10 classes with 5 examples for each class (10-way-5-shot)'. |
| Hardware Specification | Yes | All the experiments are conducted on one NVIDIA A40 GPU with Pytorch. |
| Software Dependencies | No | The paper mentions 'Pytorch' but does not specify a version number for it or any other software dependencies. |
| Experiment Setup | Yes | In the training process, we set 50 training epochs for CUB-200-2011 in the base session and 10 epochs for each incremental session. For the CIFAR 100 and mini Image Net datasets, we set 30 training epochs in the base session and 10 epochs in each incremental session. α is set to 0.5, 2, 0.5, and β is set to 1.5, 1, and 0.5 for the three datasets, respectively. The batch size is set as 256. The learning rate is initialized as 1e-3 in the base session and 1e-4 in each incremental session, which all adopt the cosine annealing schedule. Adam W [Kingma and Ba, 2014] is utilized as the model optimizer. |