# DeepWisdom/AutoDL

Automated Deep Learning without ANY human intervention. 1'st Solution for AutoDL challenge@NeurIPS.

Repository: https://github.com/DeepWisdom/AutoDL
Canonical: https://ross.abutalabs.com/products/autodl
Homepage: http://fuzhi.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: autodl, automl, nas, feature-engineering, model-selection, full-automl, artificial-intelligence, lightgbm, resnet, pytorch, tensorflow, python, autodl-challenge, ai, deeplearning, data-science, machine-learning, big-data, multi-label, automated-machine-learning
Last push: 2022-09-23T22:40:53+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2344, "days_push": 1440, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1196, forks 214 (observed 2026-08-28T04:03:56.962936+00:00)

## What it is
AutoDL is a fully automated deep learning framework from DeepWisdom that won the NeurIPS AutoDL challenge. It performs automatic feature engineering, model selection, and hyperparameter tuning for multi-label classification across image, video, audio, text, and tabular data with no human intervention.

## Use cases
- automatically classify images without writing model code
- build a text classifier with zero manual tuning
- train a tabular data classifier automatically
- classify audio or video datasets hands-free
- run automated machine learning on small datasets
- compare automated models across multiple modalities

## When to choose
- you need fully automated classification across any data modality
- you lack ML expertise and want no manual feature or model tuning
- you need fast baseline models in seconds
- you want a proven competition-winning AutoML pipeline

## When to avoid
- you need cutting-edge, actively maintained AutoML tooling
- your task is regression or unsupervised learning rather than classification
- you require fine-grained manual control over architectures
- you need support for recent Python or CUDA versions

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, data-science
- domain: machine-learning, deep-learning, artificial-intelligence, data-science
- platform: python
- tags: automl, nas, neural-architecture-search, feature-engineering, model-selection, multi-label-classification, autodl-challenge, lightgbm, pytorch, tensorflow, linux, gpu

## Member repositories
- DeepWisdom/AutoDL (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:56.962936+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:22:00.574346+00:00, confidence not recorded.
  - readme: https://github.com/DeepWisdom/AutoDL (fetched 2026-08-28T04:03:56.962936+00:00, sha 33555cf9fc41)
  - homepage: http://fuzhi.ai (fetched 2026-08-29T12:29:12.502224+00:00, sha 12e13cc74003)
  - site_page: https://atoms.dev/about (fetched 2026-08-29T12:29:12.511247+00:00, sha 655612249838)
  - site_page: https://atoms.dev/pricing (fetched 2026-08-29T12:29:12.513137+00:00, sha 0236aa8148c6)
- Data as of 2026-08-30T08:39:29.467469+00:00.