C3Box: A CLIP-based Class-Incremental Learning Toolbox

TL;DR

Problem: catastrophic forgetting in Class-Incremental Learning (CIL) is a core challenge in evolving data distributions. Approach: C3Box, a modular toolbox that unifies traditional, ViT-based, and CLIP-based CIL methods within a JSON-configured, PyCIL-inspired framework. Contributions: integration of 17 methods, standardized evaluation metrics including $A_B$ (last accuracy), $\bar{A}$ (average accuracy), and $F_B$ (forgetting), and cross‑platform, reproducible execution using OpenCLIP and common libraries. Findings: preliminary experiments on Aircraft and CIFAR100 show CLIP-based CIL methods outperform traditional baselines, demonstrating CLIP’s robustness to forgetting. Impact: C3Box provides a practical, extensible benchmark platform to drive fair comparisons and rapid adoption of CLIP-based continual learning research.

Abstract

Traditional machine learning systems are typically designed for static data distributions, which suffer from catastrophic forgetting when learning from evolving data streams. Class-Incremental Learning (CIL) addresses this challenge by enabling learning systems to continuously learn new classes while preserving prior knowledge. With the rise of pre-trained models (PTMs) such as CLIP, leveraging their strong generalization and semantic alignment capabilities has become a promising direction in CIL. However, existing CLIP-based CIL methods are often scattered across disparate codebases, rely on inconsistent configurations, hindering fair comparisons, reproducibility, and practical adoption. Therefore, we propose C3Box (CLIP-based Class-inCremental learning toolBOX), a modular and comprehensive Python toolbox. C3Box integrates representative traditional CIL methods, ViT-based CIL methods, and state-of-the-art CLIP-based CIL methods into a unified CLIP-based framework. By inheriting the streamlined design of PyCIL, C3Box provides a JSON-based configuration and standardized execution pipeline. This design enables reproducible experimentation with low engineering overhead and makes C3Box a reliable benchmark platform for continual learning research. Designed to be user-friendly, C3Box relies only on widely used open-source libraries and supports major operating systems. The code is available at https://github.com/LAMDA-CL/C3Box.

Figures (2)

• Figure 1: Overview of C3Box and its main functionalities and modules.
• Figure 2: Reproduced incremental performance of different methods on Aircraft B0 Inc10.