FitSNAP Documentation

FitSNAP is a molecular machine learning package for LAMMPS.

Wanna get started ASAP? Check out our Colab Python notebook tutorial

GitHub repo: https://github.com/FitSNAP/FitSNAP

Indices and tables

• Index

• Module Index

• Search Page

---

User Guide

• 1. Introduction
  • 1.1. Overview of FitSNAP
  • 1.2. FitSNAP Components
• 2. Installation
  • 2.1. LAMMPS Installation
    • 2.1.1. MPI for LAMMPS and FitSNAP
    • 2.1.2. Python dependencies
    • 2.1.3. LAMMPS for FitSNAP
  • 2.2. FitSNAP Installation
    • 2.2.1. Install FitSNAP with latest LAMMPS
    • 2.2.2. Minimal conda install
• 3. Run FitSNAP
  • 3.1. Command-line options
  • 3.2. Input files
    • 3.2.1. [BISPECTRUM]
    • 3.2.2. [ACE]
    • 3.2.3. [CALCULATOR]
    • 3.2.4. [ESHIFT]
    • 3.2.5. [SOLVER]
    • 3.2.6. [SCRAPER]
    • 3.2.7. [PATH]
    • 3.2.8. [OUTFILE]
    • 3.2.9. [REFERENCE]
    • 3.2.10. [GROUPS]
    • 3.2.11. [EXTRAS]
    • 3.2.12. [MEMORY]
  • 3.3. FitSNAP Outputs
  • 3.4. Running as library
• 4. Linear Models
  • 4.1. Outputs
  • 4.2. Uncertainty Quantification (UQ)
  • 4.3. In development UQ solvers
    • 4.3.1. OPT
    • 4.3.2. MCMC
    • 4.3.3. MERR
    • 4.3.4. BCS
• 5. PyTorch Models
  • 5.1. Fitting Neural Network Potentials
  • 5.2. Loss Function
  • 5.3. Outputs and Error Calculation
    • 5.3.1. Error/Comparison files
    • 5.3.2. PyTorch model files
    • 5.3.3. Calculate errors on a test set
  • 5.4. Training Performance
    • 5.4.1. GPU Acceleration
• 6. ReaxFF Models
  • 6.1. Introduction
  • 6.2. ReaxFF functional form
  • 6.3. ReaxFF LAMMPS commands
  • 6.4. Fitting ReaxFF parameters
    • 6.4.1. FitSNAP-ReaxFF input script
  • 6.5. Available ReaxFF potentials
    • 6.5.1. Combustion Branch
    • 6.5.2. Independent Branch
    • 6.5.3. Water Branch
  • 6.6. ReaxFF Bibliography
• 7. Examples
  • 7.1. Fe_Linear_NPJ2021
  • 7.2. Fe_Pytorch_NN
  • 7.3. InP_JPCA2020
  • 7.4. InP_PACE
  • 7.5. InP_VASP
  • 7.6. MSD-Fe
  • 7.7. N2_ReaxFF
  • 7.8. Ta_JAX_NN
  • 7.9. Ta_Linear_JCP2014
  • 7.10. Ta_PACE
  • 7.11. Ta_PACE_ARD
  • 7.12. Ta_PACE_PyTorch_NN
  • 7.13. Ta_PACE_RIDGE
  • 7.14. Ta_PyTorch_NN
  • 7.15. Ta_Quadratic_JCP2018
  • 7.16. Ta_XYZ
  • 7.17. Ta_pytorch_custom_NN
  • 7.18. WBe_PRB2019
  • 7.19. WBe_PyTorch_NN
  • 7.20. library

Programmer Guide

• 1. Contributing
  • 1.1. Style Guide
  • 1.2. Documenting
• 2. Executable
  • 2.1. Data & configuration extraction
  • 2.2. Modifying the output dataframe
  • 2.3. Adding new input file keywords
  • 2.4. Adding your own Calculator
  • 2.5. Adding your own Model/Solver
• 3. Library
  • 3.1. FitSnap
  • 3.2. Scraper
  • 3.3. Calculator
  • 3.4. Solver
  • 3.5. lib/
  • 3.6. tools/
• 4. Tests