mirror of https://github.com/phonopy/phono3py.git
Change default disp-distance for pypolymlp
This commit is contained in:
Atsushi Togo
parent
b8677363b2
commit
2a9fd71e53
208
doc/pypolymlp.md
208
doc/pypolymlp.md
|
|
@ -154,24 +154,22 @@ obtained by
|
|||
``````
|
||||
|
||||
|
||||
### Steps 4-7: Force constants calculation (systematic displacements in step 5)
|
||||
### Step 4: Development of MLPs
|
||||
|
||||
After developing MLPs, displacements are generated systematically considering
|
||||
crystal symmetry.
|
||||
|
||||
Having `phono3py_params.yaml`, phono3py is executed with `--pypolymlp` option,
|
||||
The `phono3py_params.yaml` file contains the training data required for
|
||||
developing polynomial MLPs when running with the `--pypolymlp` option.
|
||||
|
||||
```
|
||||
% phono3py-load phono3py_params.yaml --pypolymlp -d
|
||||
% phono3py-load --pypolymlp phono3py_params.yaml
|
||||
_ _____
|
||||
_ __ | |__ ___ _ __ ___|___ / _ __ _ _
|
||||
| '_ \| '_ \ / _ \| '_ \ / _ \ |_ \| '_ \| | | |
|
||||
| |_) | | | | (_) | | | | (_) |__) | |_) | |_| |
|
||||
| .__/|_| |_|\___/|_| |_|\___/____/| .__/ \__, |
|
||||
|_| |_| |___/
|
||||
3.11.2
|
||||
3.11.3
|
||||
|
||||
-------------------------[time 2025-01-15 16:25:17]-------------------------
|
||||
-------------------------[time 2025-01-19 18:02:40]-------------------------
|
||||
Compiled with OpenMP support (max 10 threads).
|
||||
Running in phono3py.load mode.
|
||||
Python version 3.12.6
|
||||
|
|
@ -192,6 +190,14 @@ Displacement dataset for fc3 was read from "phono3py_params.yaml".
|
|||
Pypolymlp is a generator of polynomial machine learning potentials.
|
||||
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
|
||||
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
|
||||
Parameters:
|
||||
cutoff: 8.0
|
||||
model_type: 3
|
||||
max_p: 2
|
||||
gtinv_order: 3
|
||||
gtinv_maxl: (8, 8)
|
||||
gaussian_params1: (1.0, 1.0, 1)
|
||||
gaussian_params2: (0.0, 7.0, 10)
|
||||
Developing MLPs by pypolymlp...
|
||||
Regression: cholesky decomposition ...
|
||||
- alpha: 0.001
|
||||
|
|
@ -210,29 +216,9 @@ Regression: model selection ...
|
|||
- alpha = 1.000e+01 : rmse (train, test) = 0.00002 0.00002
|
||||
MLPs were written into "phono3py.pmlp"
|
||||
------------------------------ pypolymlp end -------------------------------
|
||||
Generate displacements
|
||||
Displacement distance: 0.001
|
||||
Evaluate forces in 292 supercells by pypolymlp
|
||||
----------------------------- Force constants ------------------------------
|
||||
Computing fc3[ 1, x, x ] using numpy.linalg.pinv.
|
||||
Displacements (in Angstrom):
|
||||
[ 0.0010 0.0000 0.0000]
|
||||
[-0.0010 0.0000 0.0000]
|
||||
Computing fc3[ 33, x, x ] using numpy.linalg.pinv.
|
||||
Displacements (in Angstrom):
|
||||
[ 0.0010 0.0000 0.0000]
|
||||
[-0.0010 0.0000 0.0000]
|
||||
Expanding fc3.
|
||||
fc3 was symmetrized.
|
||||
fc2 was symmetrized.
|
||||
Max drift of fc3: 0.000000 (xxx) 0.000000 (xxx) 0.000000 (xxx)
|
||||
Max drift of fc2: 0.000000 (yy) 0.000000 (yy)
|
||||
fc3 was written into "fc3.hdf5".
|
||||
fc2 was written into "fc2.hdf5".
|
||||
----------- None of ph-ph interaction calculation was performed. -----------
|
||||
Dataset generated using MLPs was written in "phono3py_mlp_eval_dataset.yaml".
|
||||
Generate displacements (--rd or -d) for proceeding to phonon calculations.
|
||||
Summary of calculation was written in "phono3py.yaml".
|
||||
-------------------------[time 2025-01-15 16:26:37]-------------------------
|
||||
-------------------------[time 2025-01-19 18:03:44]-------------------------
|
||||
_
|
||||
___ _ __ __| |
|
||||
/ _ \ '_ \ / _` |
|
||||
|
|
@ -246,40 +232,33 @@ saved in the `phono3py.pmlp` file, which can be reused in subsequent phono3py
|
|||
executions with the `--pypolymlp` option when only displacements (and no forces)
|
||||
are provided.
|
||||
|
||||
After the MLPs are developed, systematic displacements, such as those involving
|
||||
the displacement of one or two atoms in supercells, are generated with a
|
||||
displacement distance of 0.001 Angstrom. The forces for these supercells are
|
||||
then evaluated using pypolymlp. Both the generated displacements and the
|
||||
corresponding forces are stored in the `phono3py_mlp_eval_dataset` file.
|
||||
(systematic displacements)
|
||||
### Steps 5-7: Force constants calculation (systematic displacements in step 5)
|
||||
|
||||
### Steps 4-7: Force constants calculation (random displacements in step 5)
|
||||
|
||||
After developing MLPs, random displacements are generated by specifying
|
||||
{ref}`--rd <random_displacements_option>` option. To compute force constants
|
||||
with random displacements, an external force constants calculator is necessary.
|
||||
By default, symfc is used unless another force constants solver is explicitly
|
||||
specified.
|
||||
|
||||
Having `phono3py_params.yaml`, phono3py is executed with `--pypolymlp` option,
|
||||
With the `-d` option, displacements are systematically generated while taking
|
||||
crystal symmetry into account. When running with the `--pypolymlp` option, MLPs
|
||||
are read from `phono3py.pmlp` if the file exists. In this case, training data is
|
||||
no longer required, and files such as `phono3py.yaml` can be used as the input
|
||||
structure file.
|
||||
|
||||
```
|
||||
% phono3py-load --pypolymlp --rd 200 phono3py_params.yaml
|
||||
% phono3py-load --pypolymlp -d phono3py.yaml
|
||||
_ _____
|
||||
_ __ | |__ ___ _ __ ___|___ / _ __ _ _
|
||||
| '_ \| '_ \ / _ \| '_ \ / _ \ |_ \| '_ \| | | |
|
||||
| |_) | | | | (_) | | | | (_) |__) | |_) | |_| |
|
||||
| .__/|_| |_|\___/|_| |_|\___/____/| .__/ \__, |
|
||||
|_| |_| |___/
|
||||
3.11.2
|
||||
3.11.3
|
||||
|
||||
-------------------------[time 2025-01-15 19:30:47]-------------------------
|
||||
-------------------------[time 2025-01-19 18:06:46]-------------------------
|
||||
Compiled with OpenMP support (max 10 threads).
|
||||
Running in phono3py.load mode.
|
||||
Python version 3.12.3
|
||||
Python version 3.12.6
|
||||
Spglib version 2.5.0
|
||||
----------------------------- General settings -----------------------------
|
||||
HDF5 data compression filter: gzip
|
||||
Crystal structure was read from "phono3py_params.yaml".
|
||||
Crystal structure was read from "phono3py.yaml".
|
||||
Supercell (dim): [2 2 2]
|
||||
Primitive matrix:
|
||||
[0. 0.5 0.5]
|
||||
|
|
@ -287,33 +266,94 @@ Primitive matrix:
|
|||
[0.5 0.5 0. ]
|
||||
Spacegroup: Fm-3m (225)
|
||||
Use -v option to watch primitive cell, unit cell, and supercell structures.
|
||||
NAC parameters were read from "phono3py_params.yaml".
|
||||
Displacement dataset for fc3 was read from "phono3py_params.yaml".
|
||||
NAC parameters were read from "phono3py.yaml".
|
||||
----------------------------- pypolymlp start ------------------------------
|
||||
Pypolymlp is a generator of polynomial machine learning potentials.
|
||||
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
|
||||
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
|
||||
Developing MLPs by pypolymlp...
|
||||
Regression: cholesky decomposition ...
|
||||
- alpha: 0.001
|
||||
- alpha: 0.01
|
||||
- alpha: 0.1
|
||||
- alpha: 1.0
|
||||
- alpha: 10.0
|
||||
Clear training X.T @ X
|
||||
Calculate X.T @ X for test data
|
||||
Clear test X.T @ X
|
||||
Regression: model selection ...
|
||||
- alpha = 1.000e-03 : rmse (train, test) = 1.12211e+15 1.12211e+15
|
||||
- alpha = 1.000e-02 : rmse (train, test) = 1.12211e+15 1.12211e+15
|
||||
- alpha = 1.000e-01 : rmse (train, test) = 0.00002 0.00002
|
||||
- alpha = 1.000e+00 : rmse (train, test) = 0.00002 0.00002
|
||||
- alpha = 1.000e+01 : rmse (train, test) = 0.00002 0.00002
|
||||
MLPs were written into "phono3py.pmlp"
|
||||
Load MLPs from "phono3py.pmlp".
|
||||
------------------------------ pypolymlp end -------------------------------
|
||||
Generate displacements
|
||||
Displacement distance: 0.01
|
||||
Evaluate forces in 292 supercells by pypolymlp
|
||||
----------------------------- Force constants ------------------------------
|
||||
Computing fc3[ 1, x, x ] using numpy.linalg.pinv.
|
||||
Displacements (in Angstrom):
|
||||
[ 0.0100 0.0000 0.0000]
|
||||
[-0.0100 0.0000 0.0000]
|
||||
Computing fc3[ 33, x, x ] using numpy.linalg.pinv.
|
||||
Displacements (in Angstrom):
|
||||
[ 0.0100 0.0000 0.0000]
|
||||
[-0.0100 0.0000 0.0000]
|
||||
Expanding fc3.
|
||||
fc3 was symmetrized.
|
||||
fc2 was symmetrized.
|
||||
Max drift of fc3: 0.000000 (zzz) 0.000000 (zzz) 0.000000 (zzz)
|
||||
Max drift of fc2: 0.000000 (zz) 0.000000 (zz)
|
||||
fc3 was written into "fc3.hdf5".
|
||||
fc2 was written into "fc2.hdf5".
|
||||
----------- None of ph-ph interaction calculation was performed. -----------
|
||||
Dataset generated using MLPs was written in "phono3py_mlp_eval_dataset.yaml".
|
||||
Summary of calculation was written in "phono3py.yaml".
|
||||
-------------------------[time 2025-01-19 18:06:55]-------------------------
|
||||
_
|
||||
___ _ __ __| |
|
||||
/ _ \ '_ \ / _` |
|
||||
| __/ | | | (_| |
|
||||
\___|_| |_|\__,_|
|
||||
```
|
||||
|
||||
After the MLPs are read, systematic displacements, such as those involving the
|
||||
displacement of one or two atoms in supercells, are generated with a
|
||||
displacement distance of 0.01 Angstrom. The forces for these supercells are then
|
||||
evaluated using pypolymlp. Both the generated displacements and the
|
||||
corresponding forces are stored in the `phono3py_mlp_eval_dataset.yaml` file.
|
||||
|
||||
### Steps 5-7: Force constants calculation (random displacements in step 5)
|
||||
|
||||
Random displacements are generated by specifying {ref}`--rd
|
||||
<random_displacements_option>` option. To compute force constants with random
|
||||
displacements, an external force constants calculator is necessary. By default,
|
||||
symfc is used unless another force constants solver is explicitly specified.
|
||||
When running with the `--pypolymlp` option, MLPs are read from `phono3py.pmlp`
|
||||
if the file exists. In this case, training data is no longer required, and files
|
||||
such as `phono3py.yaml` can be used as the input structure file.
|
||||
|
||||
```
|
||||
% phono3py-load --pypolymlp --rd 200 phono3py.yaml
|
||||
_ _____
|
||||
_ __ | |__ ___ _ __ ___|___ / _ __ _ _
|
||||
| '_ \| '_ \ / _ \| '_ \ / _ \ |_ \| '_ \| | | |
|
||||
| |_) | | | | (_) | | | | (_) |__) | |_) | |_| |
|
||||
| .__/|_| |_|\___/|_| |_|\___/____/| .__/ \__, |
|
||||
|_| |_| |___/
|
||||
3.11.3
|
||||
|
||||
-------------------------[time 2025-01-19 18:13:21]-------------------------
|
||||
Compiled with OpenMP support (max 10 threads).
|
||||
Running in phono3py.load mode.
|
||||
Python version 3.12.6
|
||||
Spglib version 2.5.0
|
||||
----------------------------- General settings -----------------------------
|
||||
HDF5 data compression filter: gzip
|
||||
Crystal structure was read from "phono3py.yaml".
|
||||
Supercell (dim): [2 2 2]
|
||||
Primitive matrix:
|
||||
[0. 0.5 0.5]
|
||||
[0.5 0. 0.5]
|
||||
[0.5 0.5 0. ]
|
||||
Spacegroup: Fm-3m (225)
|
||||
Use -v option to watch primitive cell, unit cell, and supercell structures.
|
||||
NAC parameters were read from "phono3py.yaml".
|
||||
----------------------------- pypolymlp start ------------------------------
|
||||
Pypolymlp is a generator of polynomial machine learning potentials.
|
||||
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
|
||||
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
|
||||
Load MLPs from "phono3py.pmlp".
|
||||
------------------------------ pypolymlp end -------------------------------
|
||||
Generate random displacements
|
||||
Twice of number of snapshots will be generated for plus-minus displacements.
|
||||
Displacement distance: 0.001
|
||||
Displacement distance: 0.01
|
||||
Evaluate forces in 400 supercells by pypolymlp
|
||||
----------------------------- Force constants ------------------------------
|
||||
Symfc will be used to handle general (or random) displacements.
|
||||
|
|
@ -338,7 +378,7 @@ fc2 was written into "fc2.hdf5".
|
|||
----------- None of ph-ph interaction calculation was performed. -----------
|
||||
Dataset generated using MLPs was written in "phono3py_mlp_eval_dataset.yaml".
|
||||
Summary of calculation was written in "phono3py.yaml".
|
||||
-------------------------[time 2025-01-15 19:32:04]-------------------------
|
||||
-------------------------[time 2025-01-19 18:13:34]-------------------------
|
||||
_
|
||||
___ _ __ __| |
|
||||
/ _ \ '_ \ / _` |
|
||||
|
|
@ -346,20 +386,18 @@ Summary of calculation was written in "phono3py.yaml".
|
|||
\___|_| |_|\__,_|
|
||||
```
|
||||
|
||||
The development of MLPs follows the same procedure as described for the
|
||||
systematic displacements (in step 5) above.
|
||||
|
||||
After the MLPs are developed, 200 supercells with random directional
|
||||
displacements are generated. These displacements are then inverted, resulting in
|
||||
an additional 200 supercells. In total, 400 supercells are created. The forces
|
||||
for these supercells are then evaluated. Finally, the force constants are
|
||||
calculated using symfc.
|
||||
After the MLPs are read, 200 supercells with random directional displacements
|
||||
are generated. These displacements are then inverted (such as $\Delta
|
||||
\mathbf{u}_i$ and $-\Delta \mathbf{u}_i$ of all atoms $i$ in each supercell),
|
||||
resulting in an additional 200 supercells. In total, 400 supercells are created.
|
||||
The forces for these supercells are then evaluated. Finally, the force constants
|
||||
are calculated using symfc.
|
||||
|
||||
### Command options for force constants calculation
|
||||
|
||||
After obtaining the MLPs, displacements are generated using those MLPs, and the
|
||||
resulting forces are computed accordingly. The displacement distance is set by
|
||||
the `--amplitude` option, whose default value is 0.001 Angstrom. When the `--rd`
|
||||
the `--amplitude` option, whose default value is 0.01 Angstrom. When the `--rd`
|
||||
option is used, it specifies the number of supercells with random directional
|
||||
displacements. Note that to achieve accurate force constants, the actual number
|
||||
of generated supercells is twice the specified number. If `--rd` is omitted,
|
||||
|
|
@ -371,19 +409,19 @@ files, `phono3py-load` will detect `phono3py.pmlp` and then compute the force
|
|||
constants by loading the MLPs from `phono3py.pmlp` as follows:
|
||||
|
||||
```
|
||||
% phono3py-load --pypolymlp --rd 100 --amplitude 0.01 phono3py.yaml
|
||||
% phono3py-load --pypolymlp --rd 100 --amplitude 0.005 phono3py.yaml
|
||||
_ _____
|
||||
_ __ | |__ ___ _ __ ___|___ / _ __ _ _
|
||||
| '_ \| '_ \ / _ \| '_ \ / _ \ |_ \| '_ \| | | |
|
||||
| |_) | | | | (_) | | | | (_) |__) | |_) | |_| |
|
||||
| .__/|_| |_|\___/|_| |_|\___/____/| .__/ \__, |
|
||||
|_| |_| |___/
|
||||
3.11.2
|
||||
3.11.3
|
||||
|
||||
-------------------------[time 2025-01-15 19:28:22]-------------------------
|
||||
-------------------------[time 2025-01-19 18:19:02]-------------------------
|
||||
Compiled with OpenMP support (max 10 threads).
|
||||
Running in phono3py.load mode.
|
||||
Python version 3.12.3
|
||||
Python version 3.12.6
|
||||
Spglib version 2.5.0
|
||||
----------------------------- General settings -----------------------------
|
||||
HDF5 data compression filter: gzip
|
||||
|
|
@ -404,7 +442,7 @@ Load MLPs from "phono3py.pmlp".
|
|||
------------------------------ pypolymlp end -------------------------------
|
||||
Generate random displacements
|
||||
Twice of number of snapshots will be generated for plus-minus displacements.
|
||||
Displacement distance: 0.01
|
||||
Displacement distance: 0.005
|
||||
Evaluate forces in 200 supercells by pypolymlp
|
||||
----------------------------- Force constants ------------------------------
|
||||
Symfc will be used to handle general (or random) displacements.
|
||||
|
|
@ -429,7 +467,7 @@ fc2 was written into "fc2.hdf5".
|
|||
----------- None of ph-ph interaction calculation was performed. -----------
|
||||
Dataset generated using MLPs was written in "phono3py_mlp_eval_dataset.yaml".
|
||||
Summary of calculation was written in "phono3py.yaml".
|
||||
-------------------------[time 2025-01-15 19:28:30]-------------------------
|
||||
-------------------------[time 2025-01-19 18:19:09]-------------------------
|
||||
_
|
||||
___ _ __ __| |
|
||||
/ _ \ '_ \ / _` |
|
||||
|
|
|
|||
|
|
@ -2272,7 +2272,7 @@ class Phono3py:
|
|||
|
||||
The displacements may be generated by the produce_force_constants method
|
||||
with number_of_snapshots > 0. With MLP, a small distance parameter, such
|
||||
as 0.001, can be numerically stable for the computation of force
|
||||
as 0.01, can be numerically stable for the computation of force
|
||||
constants.
|
||||
|
||||
"""
|
||||
|
|
@ -2336,7 +2336,7 @@ class Phono3py:
|
|||
|
||||
The displacements may be generated by the produce_force_constants method
|
||||
with number_of_snapshots > 0. With MLP, a small distance parameter, such
|
||||
as 0.001, can be numerically stable for the computation of force
|
||||
as 0.01, can be numerically stable for the computation of force
|
||||
constants.
|
||||
|
||||
"""
|
||||
|
|
|
|||
|
|
@ -561,7 +561,7 @@ def run_pypolymlp_to_compute_forces(
|
|||
|
||||
if prepare_dataset:
|
||||
if displacement_distance is None:
|
||||
_displacement_distance = 0.001
|
||||
_displacement_distance = 0.01
|
||||
else:
|
||||
_displacement_distance = displacement_distance
|
||||
|
||||
|
|
@ -629,7 +629,7 @@ def run_pypolymlp_to_compute_phonon_forces(
|
|||
print("-" * 30 + " pypolymlp end " + "-" * 31, flush=True)
|
||||
|
||||
if displacement_distance is None:
|
||||
_displacement_distance = 0.001
|
||||
_displacement_distance = 0.01
|
||||
else:
|
||||
_displacement_distance = displacement_distance
|
||||
if log_level:
|
||||
|
|
|
|||
Loading…
Reference in New Issue