1.简况
Quick Prototyping（快速原型设计）
Build machine learning solutions on raw data（原始数据） in a few lines of code.
State-of-the-art Techniques（最先进的技术）
Automatically utilize SOTA models without expert knowledge.
Easy to Deploy（易部署）
Move from experimentation to production with cloud predictors and pre-built containers.（云预测器和预构建的容器）
Customizable（可定制）
Extensible with custom feature processing, models, and metrics.（可扩展的自定义的特征提取，模型和度量）2.
2.快速开始
Tabular Data（扁平数据）
Tabular data is data that is stored in rows and columns (hence the name “tabular” i.e. in tables), either in a flat file or a database, and is usually comprised of simple alphanumeric values. CSV/TSV, JSON, XLS(X), and XML are some of the more common formats you’ll find tabular data in, though unfortunately it does still often appear in non machine-readable formats like PDF and DOC and most first be extracted and cleaned before being used.【表格数据是存储在平面文件或数据库中的行和列中的数据(因此称为“表格”，即在表中)，通常由简单的字母数字值组成。CSV/TSV、JSON、XLS(X)和XML是一些更常见的表格数据格式，但不幸的是，它仍然经常以PDF和DOC等非机器可读格式出现，并且大多数在使用前首先被提取和清理。】
Predict the class column in a data table:
from autogluon.tabular import TabularDataset, TabularPredictor
data_root = 'https://autogluon.s3.amazonaws.com/datasets/Inc/'
train_data = TabularDataset(data_root + 'train.csv')
test_data = TabularDataset(data_root + 'test.csv')
predictor = TabularPredictor(label='class').fit(train_data=train_data)
predictions = predictor.predict(test_data)
Multimodal
其中object Detection用的是mmdet。
Time Series
from autogluon.timeseries import TimeSeriesDataFrame, TimeSeriesPredictor
data = TimeSeriesDataFrame('https://autogluon.s3.amazonaws.com/datasets/timeseries/m4_hourly/train.csv')
predictor = TimeSeriesPredictor(target='target', prediction_length=48).fit(data)
predictions = predictor.predict(data)
3.安装
Install AutoGluon using pip:
pip install autogluon
AutoGluon supports Linux, MacOS, and Windows. See Installing AutoGluon for detailed instructions.
