Save and load a Forecaster

Introduction

Once we have trained a forecaster, we usually need to save the model and load it later for further training, either model evaluation or model deployment. In this guidance, we demonstrate how to save and load a forecaster in detail.

We will take TCNForecaster and nyc_taxi dataset as an example in this guide.

Setup

Before we begin, we need to install chronos if it isn’t already available, we choose to use pytorch as deep learning backend.

!pip install --pre --upgrade bigdl-chronos[pytorch]
!pip uninstall -y torchtext # uninstall torchtext to avoid version conflict

Forecaster preparation

Before saving a forecaster, a forecaster should be created and trained. The training process is introduced in the previous guidance Train forcaster on single node in detail, therefore we directly create and train a TCNForecaster based on the nyc taxi dataset.

# get TSDataset for training and testing
tsdata_train, tsdata_test = get_data()
# get a trained forecaster
forecaster = get_trained_forecaster(tsdata_train)

Save and Load

After you have trained a forecaster, you can simply save your forecaster by calling save() with a filename. Then you should load it with the same filename.

forecaster.save("./forecaster.txt")
forecaster.load("./forecaster.txt")

If you are in a new session, then you should define a forecaster first, then load it by filename.

from bigdl.chronos.forecaster.tcn_forecaster import TCNForecaster
forecaster = TCNForecaster(past_seq_len=48,
                           future_seq_len=1,
                           input_feature_num=1,
                           output_feature_num=1)
forecaster.load("./forecaster.txt")