gluonts

We have hosted the application gluonts in order to run this application in our online workstations with Wine or directly.

Quick description about gluonts:

GluonTS is a Python package for probabilistic time series modeling, focusing on deep learning based models. GluonTS requires Python 3.6 or newer, and the easiest way to install it is via pip. We train a DeepAR-model and make predictions using the simple "airpassengers" dataset. The dataset consists of a single time-series, containing monthly international passengers between the years 1949 and 1960, a total of 144 values (12 years * 12 months). We split the dataset into train and test parts, by removing the last three years (36 months) from the train data. Thus, we will train a model on just the first nine years of data. Python has the notion of extras � dependencies that can be optionally installed to unlock certain features of a package. We make extensive use of optional dependencies in GluonTS to keep the amount of required dependencies minimal. To still allow users to opt-in to certain features, we expose many extra dependencies.

Features:

• Models written using PyTorch are available via the gluonts.torch subpackage
• MXNet based models require a version of mxnet to be installed
• GluonTS includes a thin wrapper for calling the R forecast package
• GluonTS support Parquet files using PyArrow
• The shell module offers integration with Amazon SageMaker
• One core idea in GluonTS is that we don�t produce simple values as forecasts, but actually predict distributions

Programming Language: Python.
Categories: