Skip to content

doikov/dynamic-accuracies

master
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.

dynamic-accuracies

To plot the graphs from the paper "Inexact Tensor Methods with Dynamic Accuracies" by N. Doikov and Yu. Nesterov (https://arxiv.org/abs/2002.09403) use the following commands.

The results will be placed into folder "plots/*".

Figure 1:

$ python3 run_experiment_logreg.py (It requires to download the files "data/mushrooms.txt", "data/w8a.txt", "data/a8a.txt" from https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/ in advance).

Figure 2:

$ python3 run_experiment_logsumexp.py

Figures 3, 4:

$ python3 run_experiment_logsumexp_exact.py

Figure 5:

$ python3 run_experiment_logreg_exact.py (It requires to download the files "data/mushrooms.txt", "data/w8a.txt", "data/a8a.txt", "data/phishing.txt", "data/splice.txt" from https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/ in advance).

Figure 6:

$ python3 run_experiment_averaging.py

About

Experiments with second- and third-order inexact Tensor Methods with dynamic inner accuracies.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages