Skip to content

Source code for Automatic Differentiation for the Direct Minimization Approach to the Hartree-Fock Method

License

Notifications You must be signed in to change notification settings

n-yoshikawa/automatic-differentiation-SCF

Repository files navigation

Automatic Differentiation for the Direct Minimization Approach to the Hartree-Fock Method

Source code for Automatic Differentiation for the Direct Minimization Approach to the Hartree-Fock Method (arXiv).

Dependencies

  • JAX
  • PySCF
  • SciPy
  • NumPy
  • matplotlib

You can install dependencies with requirements.txt.

pip install -r requirements.txt

How to reproduce Figures/Tables

We tested our program with Python 3.8.10 on Ubuntu 20.04.5 LTS.

Figure 2(a)

python diatomic.py H2 STO-3G

Figure 2(b)

python diatomic.py H2 3-21G

Figure 3(a)

python diatomic.py HF STO-3G

Figure 3(b)

python diatomic.py HF STO-3G

Figure 4(a)

python polyatomic_angle.py H2O

Figure 4(b)

python polyatomic_angle.py NH3

Figure 5

python ad_vs_fd.py NH3 cc-pVDZ --plot

Figure 6

python plot_step_size.py

data/step_size_h2o.csv was created by running ad_vs_fd.py in different step sizes (change line 173).

Table 3

You can calculate each row using polyatomic.py. For example,

python polyatomic.py H2O STO-3G

Table 4

You can calculate each row using ad_vs_fd.py. For example,

python ad_vs_fd.py H2O STO-3G

About

Source code for Automatic Differentiation for the Direct Minimization Approach to the Hartree-Fock Method

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages