Frequently Asked Questions

How many oxide components are available in PyGGi?

In PyGGi, thirty four oxide components are available.

Should the compositions be in wt% or mol%?

All the compositions should be entered in mol%. There is no provision to enter the compositions in wt%.

Can the proportion of oxides for multiple glass compositions be uploaded through an Excel/CSV file?

Yes, the mol% of oxide components for multiple glass compositions can be imported from an Excel/CSV file. The user can download a template by choosing the required oxides using the ‘Import’ option available on the top-left corner of the PyGGi Seer software window. The same file can then be uploaded back after filling in the required compositions. For a detailed description, have a look at the documentation here.

Is it possible to save the predicted results?

Yes, the results of the computations can be accessed later by the user through their profile window.

Is the software compatible with Linux?

Presently, PyGGi Seer and PyGGi Seerpro are available for Windows and MacOS only. Linux version may be provided in the future based on user requests. For the detailed hardware and software requirements, see Download PyGGi Seer

Which models are used for predicting the properties of glasses?

Glass properties are predicted using models trained by machine learning. Two different types of models, one trained by neural network and other trained by Gaussian process regression, are used for prediction. Further details of each model are listed in the documentation.

What is the data source for PyGGi Bank?

The experimental data in PyGGi Bank has been obtained from various literature and the open-source glass database SciGlass. The GitHub repository of the database can be found here.

What are the optimization methods used in PyGGi Zen for glass discovery?

There are four algorithms available for doing optimization:

  • Gradient descent
  • Particle swarm optimization
  • Genetic algorithm
  • Ant colony optimization

What machine learning models are used in PyGGi Zen for optimization?

Neural Network models of PyGGi Seer are used in PyGGi Zen.