AutoSkill is a comprehensive framework for extracting, managing, and evolving reusable 'skills' from conversational logs and documents. It operates as a modular application with multiple entry points: a web UI for interactive chat, a CLI for batch processing, and an OpenAI-compatible proxy server that injects relevant skills into LLM context. The system employs LLM-driven and heuristic pipelines to extract skill candidates, deduplicate and merge them, and persist them in a skill store with vector and BM25 retrieval. It also includes a self-evolution subsystem (SkillEvo) that replays experiences, evaluates performance, and mutates skills for improvement. The codebase is hybrid, containing both application components (web server, proxy, libraries) and data/research pipelines (document ingestion, trajectory extraction), but the dominant character is an application SDK designed for integration with agentic workflows. Notable craftsmanship is seen in standalone utilities like a CJK-aware text unit counter, an AST-based prompt patcher, a persistent BM25 index, and a LibreOffice socket shim, though the repo also suffers from significant code duplication (e.g., identical validators across docx/pptx/xlsx) and a large surface area of environment variables and external dependencies.
This repository is an 'awesome list' curated by the TeleAI-UAGI team, aggregating open-source and closed-source products, papers, benchmarks, tutorials, and surveys on memory systems for LLM/MLLM agents. It does not contain application business logic or runnable services; its primary output is a meticulously structured README.md that serves as a reference catalog. The accompanying Python scripts automate two maintenance tasks: (1) checking that open-source entries are ordered by GitHub star count and respecting a 100-star boundary for emerging projects, and (2) generating self-hosted star-history SVG charts. The scripts are well-crafted, with robust parsing, GitHub API integration, and careful edge-case handling (grace bands, orphan detection), making them reusable for other awesome-list maintainers. The repository's value lies in its curated content rather than its code, but the supporting infrastructure demonstrates solid engineering principles.
Renate is a Python library for continual learning that provides implementations of various algorithms (Experience Replay, DER, Super-ER, etc.) to mitigate catastrophic forgetting when updating neural networks on sequential data. It features a modular architecture separating models, data handling, update strategies, memory buffers, and drift detection, enabling users to customize continual learning pipelines. The framework supports local execution and cloud deployment via Amazon SageMaker, includes hyperparameter optimization through Syne Tune, and offers benchmarking tools for evaluating continual learning methods across diverse datasets and scenarios. Key reusable components include a memory buffer system with reservoir and class-balancing strategies, Maximum Mean Discrepancy (MMD) based drift detectors for monitoring data distribution shifts, and a composable loss component system for constructing complex continual learning objectives. The library integrates with PyTorch Lightning for training and provides abstract base classes for models and data modules to facilitate extension. Although the project is explicitly archived and no longer maintained, the codebase demonstrates solid engineering practices with well-defined interfaces and decoupled components. It has been tested on vision and NLP benchmarks (CIFAR, CLEAR, DomainNet, Hugging Face datasets) and includes synthetic scenario generators for controlled experimentation. The archival status indicates that users should expect no further updates or security patches, but the core functionality remains usable for continual learning research and applications.
This repository is a curated collection of machine unlearning research, combining a comprehensive survey paper (in the form of a Kaggle notebook) with executable implementations of several unlearning algorithms across different domains. It includes code for graph neural network unlearning (GraphProjector), federated learning unlearning (FedEraser), and approximate nearest neighbor search with unlearning (UnfoldedSelf-ReconstructionHashing, a fork of ann-benchmarks). The repo serves both as a literature index (taxonomy, evaluation metrics, datasets) and a benchmark hub for reproducing and comparing unlearning methods. While the survey is actively maintained, the algorithm codebases are research prototypes with varying levels of polish and portability.
This repository is a curated list (awesome list) of research papers, surveys, workshops, and challenges on incremental/continual learning. It contains no executable code, datasets, or reproducible pipelines. The sole content is a README.md file organizing references by year and topic, with links to papers and code repositories. While useful as a bibliography for researchers, it provides no software artifacts, data transformations, or analytical models. It cannot be used as a tool or data pipeline.