Computational Performance

⚠️ This page is a work in progress.

This page contains some example computational performance statistics for TemporAI tasks. It will be expanded we test more models, datasets, and hardware.

Measurement details

  • TemporAI version: 0.0.3

  • Last tested: Nov 2023

  • Hardware:

    • CPU: 10-core Intel Core i9-10900K

    • RAM: 64 GB

    • GPU: NVIDIA GeForce RTX 3090

GPU used whenever supported by a model.

Table of results

Method

Dataset

Action(s)

Execution Time (seconds)

Preprocessing : Imputation : ts_tabular_imputer (imputer=mice)

sine (n=300)

Fit and transform

5.16

Preprocessing : Imputation : ts_tabular_imputer (imputer=mean)

sine (n=300)

Fit and transform

0.57

Survival Analysis : dynamic_deephit (n_iter=100)

PBC (n=312)

Fit and predict

3.56

Treatment Effects : Temporal : crn_regressor (epochs=100)

Dummy dataset (n=300)

Fit

327.57

Prediction : One-off : nn_classifier (mode=“transformer”, n_iter=100)

Dummy dataset (n=300)

Fit and predict

0.81

More cases to be added…