--- title: "autokeras vs nni" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/keras-team-autokeras-vs-microsoft-nni" tools: ["keras-team-autokeras", "microsoft-nni"] --- # autokeras vs nni *GraphCanon updated Aug 4, 2026* ## Verdict Pick autokeras if autoKeras simplifies deep learning model design through automated neural architecture search and is compatible with Python 3.7+ and TensorFlow 2.8.0+; pick nni if nNI is an AutoML toolkit that supports feature engineering, neural architecture search, model compression, and hyperparameter tuning with the flexibility of Python programming. [autokeras](http://autokeras.com/) reports 9.3k GitHub stars, 1.4k forks, and 161 open issues, last pushed Nov 25, 2025. [nni](https://nni.readthedocs.io) has 14k stars, 1.9k forks, and 415 open issues, last pushed Jul 3, 2024. Figures are from public GitHub metadata via [autokeras's repository](https://github.com/keras-team/autokeras) and [nni's repository](https://github.com/microsoft/nni). | | [autokeras](/tools/keras-team-autokeras.md) | [nni](/tools/microsoft-nni.md) | | --- | --- | --- | | Tagline | AutoML library for deep learning | An open source AutoML toolkit for automating machine learning lifecycle | | Stars | 9,328 | 14,363 | | Forks | 1,393 | 1,853 | | Open issues | 161 | 415 | | Language | Python | Python | | Adopt for | AutoKeras simplifies deep learning model design through automated neural architecture search and is compatible with Python 3.7+ and TensorFlow 2.8.0+. | NNI is an AutoML toolkit that supports feature engineering, neural architecture search, model compression, and hyperparameter tuning with the flexibility of Python programming. | | Persona | - | - | | Runtime | - | - | | License | Apache-2.0 | MIT | | Categories | Developer Tools, Model Training | Model Training | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [autokeras](/tools/keras-team-autokeras.md) | [nni](/tools/microsoft-nni.md) | | --- | --- | --- | | Maintenance | Slowing (36%) | Archived (8%) | | Days since push | 251d | 762d | | Archived on GitHub | No | Yes | | Open issues (now) | 161 | 415 | | Full report | [trust report](/tools/keras-team-autokeras/trust.md) | [trust report](/tools/microsoft-nni/trust.md) | ## Shared compatibility - **Python**: [autokeras](/tools/keras-team-autokeras.md) - Python runtime; [nni](/tools/microsoft-nni.md) - Python runtime ## Decision facts: autokeras - **Adopt for:** AutoKeras simplifies deep learning model design through automated neural architecture search and is compatible with Python 3.7+ and TensorFlow 2.8.0+. ## Decision facts: nni - **Adopt for:** NNI is an AutoML toolkit that supports feature engineering, neural architecture search, model compression, and hyperparameter tuning with the flexibility of Python programming. ## Choose when ### Choose autokeras if… - License: autokeras is Apache-2.0, nni is MIT. - Tags unique to autokeras: autodl, keras, machine-learning, neural-architecture-search. - Also covers Developer Tools. - When your project involves deep learning tasks requiring minimal manual intervention in designing models. ### Choose nni if… - License: nni is MIT, autokeras is Apache-2.0. - Tags unique to nni: automated-machine-learning, bayesian-optimization, data-science, deep-neural-network. - nni ships Docker support for self-hosted deployment. - You need to automate extensive parts of your machine learning lifecycle from preprocessing to deployment. ## When NOT to use autokeras - When working with Python versions older than 3.7 or TensorFlow versions older than 2.8.0, as AutoKeras is not compatible. - If your project emphasizes transparent, understandable model architecture over automated generation without human oversight. ## When NOT to use nni - You require real-time automated tuning capabilities, as NNI focuses on batch processing and model training scenarios. - If your project demands direct integration with specific deep learning frameworks beyond PyTorch and TensorFlow, NNI support is limited to these two environments. ## Common questions ### What is the difference between autokeras and nni? autokeras: AutoML library for deep learning. nni: An open source AutoML toolkit for automating machine learning lifecycle. See the comparison table for live GitHub stats and shared categories. ### When should I choose autokeras over nni? Choose autokeras over nni when License: autokeras is Apache-2.0, nni is MIT; Tags unique to autokeras: autodl, keras, machine-learning, neural-architecture-search; Also covers Developer Tools; When your project involves deep learning tasks requiring minimal manual intervention in designing models. ### When should I choose nni over autokeras? Choose nni over autokeras when License: nni is MIT, autokeras is Apache-2.0; Tags unique to nni: automated-machine-learning, bayesian-optimization, data-science, deep-neural-network; nni ships Docker support for self-hosted deployment; You need to automate extensive parts of your machine learning lifecycle from preprocessing to deployment. ### When should I avoid autokeras? When working with Python versions older than 3.7 or TensorFlow versions older than 2.8.0, as AutoKeras is not compatible. If your project emphasizes transparent, understandable model architecture over automated generation without human oversight. ### When should I avoid nni? You require real-time automated tuning capabilities, as NNI focuses on batch processing and model training scenarios. If your project demands direct integration with specific deep learning frameworks beyond PyTorch and TensorFlow, NNI support is limited to these two environments. ### Is autokeras or nni more popular on GitHub? nni has more GitHub stars (14,363 vs 9,328). Stars measure visibility, not whether either tool fits your constraints. ### Are autokeras and nni open source? Yes - both are open-source projects on GitHub (autokeras: Apache-2.0, nni: MIT). ### Where can I find alternatives to autokeras or nni? GraphCanon lists graph-backed alternatives at [autokeras alternatives](/tools/keras-team-autokeras/alternatives) and [nni alternatives](/tools/microsoft-nni/alternatives) ([autokeras markdown twin](/tools/keras-team-autokeras/alternatives.md), [nni markdown twin](/tools/microsoft-nni/alternatives.md)), ranked by typed relationship edges rather than popularity votes. ### Is there a machine-readable version of this comparison? Yes. The markdown twin at [this comparison](/compare/keras-team-autokeras-vs-microsoft-nni.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, autokeras or nni? autokeras: Slowing. nni: Archived. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance. ### Where are the full trust reports for autokeras and nni? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [autokeras trust report](/tools/keras-team-autokeras/trust); [nni trust report](/tools/microsoft-nni/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=keras-team-autokeras`](/api/graphcanon/graph?tool=keras-team-autokeras) - LLM index: [/llms.txt](/llms.txt) - Full corpus: [/llms-full.txt](/llms-full.txt) _GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_