pytorch-ident

System Identification with PyTorch.

pytorch-ident is an open-source python library for system identification with deep learning tools based on the PyTorch framework. The project is hosted on this GitHub repository.

Requirements

In order to run pytorch-ident, you need a python 3.x environment and the following packages:

• PyTorch

• numpy

• scipy

• matplotlib

Installation

1. Copy or clone the pytorch-ident project into a local folder. For instance, run

git clone https://github.com/forgi86/pytorch-ident.git

from the command line

2. Navigate to your local pytorch-ident folder

cd LOCAL_FOLDER

where LOCAL_FOLDER is the folder where you have just downloaded the code in step 2

3. Install pytorch-ident in your python environment: run

pip install .

from the command line, in the working folder LOCAL_FOLDER

Getting started

The best way to get started with pytorch-ident is to run the examples.

API Documentation

• Datasets
  • SubsequenceDataset
• Discrete-time state-space identification
  • Discrete-time models
  • Discrete-time simulators
• Continuous-time state-space identification
  • Continuous-time models
  • Continuous-time simulators
• dynoNet
  • Linear dynamical blocks
  • Static non-linear blocks
• Performance metrics
  • r_squared()
  • rmse()
  • nrmse()
  • error_mean()
  • mae()
  • fit_index()

Indices and tables

• Index

• Module Index

• Search Page