DeepWisdom / Autodl
Licence: apache-2.0
Automated Deep Learning without ANY human intervention. 1'st Solution for AutoDL [email protected]
Stars: ✭ 854
Programming Languages
python
139335 projects - #7 most used programming language
Labels
Projects that are alternatives of or similar to Autodl
Auto ml
[UNMAINTAINED] Automated machine learning for analytics & production
Stars: ✭ 1,559 (+82.55%)
Mutual labels: artificial-intelligence, data-science, deeplearning, automl, feature-engineering, lightgbm, automated-machine-learning
Nni
An open source AutoML toolkit for automate machine learning lifecycle, including feature engineering, neural architecture search, model compression and hyper-parameter tuning.
Stars: ✭ 10,698 (+1152.69%)
Mutual labels: data-science, nas, automl, feature-engineering, automated-machine-learning
Mljar Supervised
Automated Machine Learning Pipeline with Feature Engineering and Hyper-Parameters Tuning 🚀
Stars: ✭ 961 (+12.53%)
Mutual labels: data-science, automl, feature-engineering, lightgbm, automated-machine-learning
Hyperparameter hunter
Easy hyperparameter optimization and automatic result saving across machine learning algorithms and libraries
Stars: ✭ 648 (-24.12%)
Mutual labels: artificial-intelligence, ai, data-science, feature-engineering, lightgbm
Mlbox
MLBox is a powerful Automated Machine Learning python library.
Stars: ✭ 1,199 (+40.4%)
Mutual labels: data-science, automl, lightgbm, automated-machine-learning
Tpot
A Python Automated Machine Learning tool that optimizes machine learning pipelines using genetic programming.
Stars: ✭ 8,378 (+881.03%)
Mutual labels: data-science, automl, feature-engineering, automated-machine-learning
Lightautoml
LAMA - automatic model creation framework
Stars: ✭ 196 (-77.05%)
Mutual labels: data-science, automl, feature-engineering, automated-machine-learning
Pycm
Multi-class confusion matrix library in Python
Stars: ✭ 1,076 (+26%)
Mutual labels: artificial-intelligence, ai, data-science, deeplearning
Blurr
Data transformations for the ML era
Stars: ✭ 96 (-88.76%)
Mutual labels: artificial-intelligence, ai, data-science, feature-engineering
Lale
Library for Semi-Automated Data Science
Stars: ✭ 198 (-76.81%)
Mutual labels: artificial-intelligence, data-science, automl, automated-machine-learning
Hyperactive
A hyperparameter optimization and data collection toolbox for convenient and fast prototyping of machine-learning models.
Stars: ✭ 182 (-78.69%)
Mutual labels: artificial-intelligence, data-science, feature-engineering, automated-machine-learning
My Data Competition Experience
本人多次机器学习与大数据竞赛Top5的经验总结，满满的干货，拿好不谢
Stars: ✭ 271 (-68.27%)
Mutual labels: data-science, automl, feature-engineering, lightgbm
Transmogrifai
TransmogrifAI (pronounced trăns-mŏgˈrə-fī) is an AutoML library for building modular, reusable, strongly typed machine learning workflows on Apache Spark with minimal hand-tuning
Stars: ✭ 2,084 (+144.03%)
Mutual labels: ai, automl, feature-engineering, automated-machine-learning
Fixy
Amacımız Türkçe NLP literatüründeki birçok farklı sorunu bir arada çözebilen, eşsiz yaklaşımlar öne süren ve literatürdeki çalışmaların eksiklerini gideren open source bir yazım destekleyicisi/denetleyicisi oluşturmak. Kullanıcıların yazdıkları metinlerdeki yazım yanlışlarını derin öğrenme yaklaşımıyla çözüp aynı zamanda metinlerde anlamsal analizi de gerçekleştirerek bu bağlamda ortaya çıkan yanlışları da fark edip düzeltebilmek.
Stars: ✭ 165 (-80.68%)
Mutual labels: artificial-intelligence, ai, data-science, deeplearning
Pba
Efficient Learning of Augmentation Policy Schedules
Stars: ✭ 461 (-46.02%)
Mutual labels: artificial-intelligence, data-science, automl, automated-machine-learning
Featuretools
An open source python library for automated feature engineering
Stars: ✭ 5,891 (+589.81%)
Mutual labels: data-science, automl, feature-engineering, automated-machine-learning
Artificio
Deep Learning Computer Vision Algorithms for Real-World Use
Stars: ✭ 326 (-61.83%)
Mutual labels: artificial-intelligence, ai, data-science
Tensorwatch
Debugging, monitoring and visualization for Python Machine Learning and Data Science
Stars: ✭ 3,191 (+273.65%)
Mutual labels: ai, data-science, deeplearning
Csinva.github.io
Slides, paper notes, class notes, blog posts, and research on ML 📉, statistics 📊, and AI 🤖.
Stars: ✭ 342 (-59.95%)
Mutual labels: artificial-intelligence, ai, data-science
Machinejs
[UNMAINTAINED] Automated machine learning- just give it a data file! Check out the production-ready version of this project at ClimbsRocks/auto_ml
Stars: ✭ 412 (-51.76%)
Mutual labels: data-science, automl, automated-machine-learning
English | 简体中文
AutoDL [email protected] 冠军方案，竞赛细节参见 AutoDL Competition。
1. AutoDL是什么？
AutoDL聚焦于自动进行任意模态（图像、视频、语音、文本、表格数据）多标签分类的通用算法，可以用一套标准算法流解决现实世界的复杂分类问题，解决调数据、特征、模型、超参等烦恼，最短10秒就可以做出性能优异的分类器。本工程在不同领域的24个离线数据集、15个线上数据集都获得了极为优异的成绩。AutoDL拥有以下特性：
☕ 全自动：全自动深度学习/机器学习框架，全流程无需人工干预。数据、特征、模型的所有细节都已调节至最佳，统一解决了资源受限、数据倾斜、小数据、特征工程、模型选型、网络结构优化、超参搜索等问题。只需要准备数据，开始AutoDL，然后喝一杯咖啡。
🌌 通用性：支持任意模态，包括图像、视频、音频、文本和结构化表格数据，支持任意多标签分类问题，包括二分类、多分类、多标签分类。它在不同领域都获得了极其优异的成绩，如行人识别、行人动作识别、人脸识别、声纹识别、音乐分类、口音分类、语言分类、情感分类、邮件分类、新闻分类、广告优化、推荐系统、搜索引擎、精准营销等等。
👍 效果出色：AutoDL竞赛获得压倒性优势的冠军方案，包含对传统机器学习模型和最新深度学习模型支持。模型库包括从LR/SVM/LGB/CGB/XGB到ResNet*/MC3/DNN/ThinResnet*/TextCNN/RCNN/GRU/BERT等优选出的冠军模型。
⚡ 极速/实时：最快只需十秒即可获得极具竞争力的模型性能。结果实时刷新（秒级），无需等待即可获得模型实时效果反馈。
2. 目录
3. 效果
-
预赛榜单（DeepWisdom总分第一，平均排名1.2，在5个数据集中取得了4项第一）
-
决赛榜单（DeepWisdom总分第一，平均排名1.8，在10个数据集中取得了7项第一）
4. AutoDL竞赛使用说明
-
基础环境
python>=3.5 CUDA 10 cuDNN 7.5 -
clone仓库
cd <path_to_your_directory> git clone https://github.com/DeepWisdom/AutoDL.git -
预训练模型准备 下载模型 speech_model.h5 放至
AutoDL_sample_code_submission/at_speech/pretrained_models/目录。 -
可选：使用与竞赛同步的docker环境
- CPU
cd path/to/autodl/ docker run -it -v "$(pwd):/app/codalab" -p 8888:8888 evariste/autodl:cpu-latest- GPU
nvidia-docker run -it -v "$(pwd):/app/codalab" -p 8888:8888 evariste/autodl:gpu-latest -
数据集准备：使用
AutoDL_sample_data中样例数据集，或批量下载竞赛公开数据集。 -
进行本地测试
python run_local_test.py
本地测试完整使用。
python run_local_test.py -dataset_dir='AutoDL_sample_data/miniciao' -code_dir='AutoDL_sample_code_submission'
您可在 AutoDL_scoring_output/ 目录中查看实时学习曲线反馈的HTML页面。
细节可参考 AutoDL Challenge official starting_kit.
4.1. 使用效果示例（横轴为对数时间轴，纵轴为AUC）
可以看出，在五个不同模态的数据集下，AutoDL算法流都获得了极为出色的全时期效果，可以在极短的时间内达到极高的精度。
5. 安装
本仓库在 Python 3.6+, PyTorch 1.3.1 和 TensorFlow 1.15上测试.
你应该在虚拟环境 中安装autodl。 如果对虚拟环境不熟悉，请看 用户指导.
用合适的Python版本创建虚拟环境，然后激活它。
5.1 windows10 安装过程
5.1.1 安装 cuda 10.0 和 cudnn v7.6.2.24
- CUDA 10.0下载
- cuDNN下载
- 百度云 提取码：xb9x
5.1.2 安装 Miniconda3-4.5.4-Windows-x86_64.exe
- Miniconda3-4.5.4-Windows-x86_64.exe
- 百度云 提取码：xb9x
5.1.3 安装 visualcppbuildtools_full.exe
- visualcppbuildtools_full.exe
- 百度云 提取码：xb9x
5.1.4 创建 start_env.bat 文件
- 将其移动到安装的
Miniconda3同级目录下
cmd.exe "/K" .\Miniconda3\Scripts\activate.bat .\Miniconda3
5.1.5 双击 start_env.bat 安装 autodl-gpu
conda install pytorch==1.3.1
conda install torchvision -c pytorch
pip install autodl-gpu
5.2 Linux安装
pip install autodl-gpu
6. 快速上手
6.1. 快速上手之AutoDL本地效果测试
指导参见 快速上手之AutoDL本地效果测试，样例代码参见 examples/run_local_test.py
6.2. 快速上手之图像分类
参见 快速上手之图像分类，样例代码参见 examples/run_image_classification_example.py
6.3. 快速上手之视频分类
指导参见 快速上手之视频分类，样例代码参见examples/run_video_classification_example.py
6.4. 快速上手之音频分类
指导参见 快速上手之音频分类，样例代码参见examples/run_speech_classification_example.py
6.5. 快速上手之文本分类
指导参见 快速上手之文本分类，样例代码参见examples/run_text_classification_example.py。
6.6. 快速上手之表格分类
指导参见 快速上手之表格分类，样例代码参见examples/run_tabular_classification_example.py.
7. 可用数据集
7.1. (可选) 下载数据集
python download_public_datasets.py
7.2. 公共数据集信息
| # | Name | Type | Domain | Size | Source | Data (w/o test labels) | Test labels |
|---|---|---|---|---|---|---|---|
| 1 | Munster | Image | HWR | 18 MB | MNIST | munster.data | munster.solution |
| 2 | City | Image | Objects | 128 MB | Cifar-10 | city.data | city.solution |
| 3 | Chucky | Image | Objects | 128 MB | Cifar-100 | chucky.data | chucky.solution |
| 4 | Pedro | Image | People | 377 MB | PA-100K | pedro.data | pedro.solution |
| 5 | Decal | Image | Aerial | 73 MB | NWPU VHR-10 | decal.data | decal.solution |
| 6 | Hammer | Image | Medical | 111 MB | Ham10000 | hammer.data | hammer.solution |
| 7 | Kreatur | Video | Action | 469 MB | KTH | kreatur.data | kreatur.solution |
| 8 | Kreatur3 | Video | Action | 588 MB | KTH | kreatur3.data | kreatur3.solution |
| 9 | Kraut | Video | Action | 1.9 GB | KTH | kraut.data | kraut.solution |
| 10 | Katze | Video | Action | 1.9 GB | KTH | katze.data | katze.solution |
| 11 | data01 | Speech | Speaker | 1.8 GB | -- | data01.data | data01.solution |
| 12 | data02 | Speech | Emotion | 53 MB | -- | data02.data | data02.solution |
| 13 | data03 | Speech | Accent | 1.8 GB | -- | data03.data | data03.solution |
| 14 | data04 | Speech | Genre | 469 MB | -- | data04.data | data04.solution |
| 15 | data05 | Speech | Language | 208 MB | -- | data05.data | data05.solution |
| 16 | O1 | Text | Comments | 828 KB | -- | O1.data | O1.solution |
| 17 | O2 | Text | Emotion | 25 MB | -- | O2.data | O2.solution |
| 18 | O3 | Text | News | 88 MB | -- | O3.data | O3.solution |
| 19 | O4 | Text | Spam | 87 MB | -- | O4.data | O4.solution |
| 20 | O5 | Text | News | 14 MB | -- | O5.data | O5.solution |
| 21 | Adult | Tabular | Census | 2 MB | Adult | adult.data | adult.solution |
| 22 | Dilbert | Tabular | -- | 162 MB | -- | dilbert.data | dilbert.solution |
| 23 | Digits | Tabular | HWR | 137 MB | MNIST | digits.data | digits.solution |
| 24 | Madeline | Tabular | -- | 2.6 MB | -- | madeline.data | madeline.solution |
8. 贡献代码
❤️ 请毫不犹豫参加贡献 Open an issue 或提交 PRs。
9. 加入社区
10. 开源协议
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].