- based on tensorflow v2.3.0
- use timeseries_dataset_from_array function which was introduced in tf v2.3.0
- This module converts time series data from dataframe type to sliding window type
- to use as input in RNN based layer
- This module was based on tensorflow official docs, just aggregate some functions and add small tuning to use it more efficiently.
- to make it possible to control batch_size, sequence_stride_size and shuffle more freely.
pip install slidingwindw_generator
from slidingwindow_generator.slidingwindow_generator import SlidingWindowGenerator
SlidingWindowGenerator(...)
- you can see detailed example with real data and RNN in
./tests/example.ipynb