Contents Menu Expand Light mode Dark mode Auto light/dark mode
Torch Spatiotemporal
Logo

Usage

  • Quickstart
  • Data Structures
  • Spatiotemporal Dataset
  • Notebooks
    • A Gentle Introduction to tsl

Package API

  • Data structures
    • Data objects
    • PyTorch Datasets
    • Organizing data
    • Preprocessing
  • Datasets
    • Prototypes
    • Datasets in tsl
  • NN
    • Layers
    • Blocks
    • Models
    • Functional
  • Inference engines
  • Metrics
    • Numpy metrics
    • PyTorch metrics
  • Experiment
  • Transforms
  • Operations

Indices

  • Index
  • Module Index
  v: latest
Versions
latest
dev
Downloads
On Read the Docs
Project Home
Builds
Back to top
Edit this page

Metrics#

The module tsl.metrics exposes API to evaluate common metrics on both ndarray and Tensor objects.

Metrics packages

  • Numpy metrics
    • mae()
    • nmae()
    • mape()
    • mse()
    • rmse()
    • nrmse()
    • nrmse_2()
    • r2()
    • mre()
  • PyTorch metrics
    • MaskedMetric
    • MaskedMAE
    • MaskedMSE
    • MaskedMRE
    • MaskedMAPE
    • MaskedPinballLoss
    • mae()
    • nmae()
    • mape()
    • mse()
    • rmse()
    • nrmse()
    • nrmse_2()
    • r2()
    • mre()
    • multi_quantile_pinball_loss()
    • MaskedMetricWrapper
    • SelectMetricWrapper
    • convert_to_masked_metric()
Next
Numpy metrics
Previous
Inference engines
Copyright © 2023, Andrea Cini, Ivan Marisca
Made with Sphinx and @pradyunsg's Furo