A curated list of tools, platforms, libraries, and resources to support AI and machine learning research — including experimentation, literature search, dataset exploration, evaluation, visualization, reproducibility, and more.

Support ongoing maintenance and curation via GitHub Sponsors.

• Experimentation & Notebooks
• Model Training & Experiments
• Evaluation & Benchmarking
• Visualization & Interpretability
• Literature Search & Paper Tools
• Dataset Tools
• Reproducibility & Experiment Tracking
• Learning Resources
• Related Awesome Lists

• JupyterLab – The de-facto environment for interactive research notebooks.
• Google Colab – Free GPU-enabled notebooks for research and prototyping.
• Kaggle Notebooks – Notebook environment with free GPUs and large datasets.
• Deepnote – Collaborative research notebooks with real-time editing.

• PyTorch – Most widely used deep learning framework in research.
• TensorFlow – Full-stack ML framework from Google.
• JAX – High-performance, functional ML framework used in cutting-edge research.
• HuggingFace Transformers – Library of pretrained models and research-ready architectures.
• Lightning – High-level framework for scalable model training.
• Fast.ai – Research-friendly training library built on PyTorch.
• DeepSpeed – Training optimizations for large models.
• Colossal-AI – Distributed training and optimization for LLMs.

• OpenAI Evals – Framework for evaluating generative models.
• HELM – Holistic evaluation across diverse reasoning and safety metrics.
• LM Evaluation Harness – Standardized benchmark suite for LLMs.
• RAGAS – Evaluation metrics for RAG systems.
• Arize Phoenix – Open-source observability and evaluation toolkit.
• W&B Evaluation Tools – Model evaluation tracking and visualization.

• TensorBoard – Visualization suite for model training and metrics.
• Weights & Biases – Experiment dashboarding and visual analytics.
• Captum – Model interpretability for PyTorch.
• LIT (Language Interpretability Tool) – Interactive model analysis from Google PAIR.
• Netron – Neural network model visualizer (ONNX, PyTorch, TensorFlow).
• Aletheia – Tools for AI interpretability and artifact inspection.

• arXiv – Primary repository for ML and AI research papers.
• Papers With Code – SOTA benchmarks and paper-to-code mappings.
• ar5iv – Effortless HTML versions of arXiv papers.
• Semantic Scholar – AI-powered academic search engine.
• Connected Papers – Visual graph exploration of related research.
• ResearchRabbit – Visual discovery tool for academic literature networks.
• Scholarcy – AI assistant for summarizing research papers.
• ExplainPaper – Highlight sections of papers and get explanations.

• HuggingFace Datasets – Huge collection of ready-to-use datasets.
• Kaggle Datasets – Public datasets across all domains.
• TensorFlow Datasets – Standardized datasets for ML research.
• Dataset Search (Google) – Search engine for public datasets.
• NVIDIA NGC Datasets – Vision, speech, and scientific datasets.
• OpenML – Platform for dataset exploration and benchmarking.

• Weights & Biases – End-to-end experiment tracking and reproducibility tools.
• MLflow – Open-source experiment tracking, models, and pipelines.
• DVC (Data Version Control) – Git-style versioning for datasets and models.
• Guild AI – Experiment management for machine learning research.
• Aim – Fast, open-source experiment tracker.
• Reproducibility Checklist – Community guidelines for reproducible ML.

• Deep Learning Book (Goodfellow) – The foundational text for modern deep learning.
• Dive into Deep Learning – Interactive, code-driven deep learning book.
• Fast.ai Courses – Application-focused deep learning education.
• Stanford CS231n – The classic computer vision and deep learning course.
• MIT Intro to Deep Learning – University course with hands-on labs.
• HuggingFace Course – Full curriculum for LLMs, transformers, and pipelines.

• Awesome AI
• Awesome Machine Learning
• Awesome AI Research Papers
• Awesome LLMOps
• Awesome Computer Vision

Contributions are welcome. Please ensure your submission fully follows the requirements outlined in CONTRIBUTING.md, including formatting, scope alignment, and category placement.

Pull requests that do not adhere to the contribution guidelines may be closed.