@@ -64,7 +64,7 @@ in the distribution from GitHub or PyPI.
64
64
` phono3py_params.yaml ` . Use {ref}` --cf3 <cf3_option> ` and {ref}`--sp
65
65
<sp_option>` option simultaneously.
66
66
4 . Develop MLPs. By default, 90 and 10 percents of the dataset are used for the
67
- training and test, respectively. At this step ` phono3py.pmlp ` is saved.
67
+ training and test, respectively. At this step ` polymlp.yaml ` is saved.
68
68
5 . Generate displacements in supercells either systematic or random displacements.
69
69
6 . Evaluate MLPs for forces of the supercells generated in step 5.
70
70
7 . Calculate force constants from displacement-force dataset from steps 5 and 6.
@@ -217,7 +217,7 @@ Regression: model selection ...
217
217
- alpha = 1.000e-01 : rmse (train, test) = 0.00002 0.00002
218
218
- alpha = 1.000e+00 : rmse (train, test) = 0.00002 0.00002
219
219
- alpha = 1.000e+01 : rmse (train, test) = 0.00002 0.00002
220
- MLPs were written into "phono3py.pmlp "
220
+ MLPs were written into "polymlp.yaml "
221
221
------------------------------ pypolymlp end -------------------------------
222
222
Generate displacements (--rd or -d) for proceeding to phonon calculations.
223
223
Summary of calculation was written in "phono3py.yaml".
@@ -231,7 +231,7 @@ Summary of calculation was written in "phono3py.yaml".
231
231
232
232
Information about the development of MLPs using pypolymlp is provided between
233
233
the ` pypolymlp start ` and ` pypolymlp end ` sections. The polynomial MLPs are
234
- saved in the ` phono3py.pmlp ` file, which can be reused in subsequent phono3py
234
+ saved in the ` polymlp.yaml ` file, which can be reused in subsequent phono3py
235
235
executions with the ` --pypolymlp ` option when only displacements (and no forces)
236
236
are provided.
237
237
@@ -240,7 +240,7 @@ are provided.
240
240
241
241
With the ` -d ` option, displacements are systematically generated while taking
242
242
crystal symmetry into account. When running with the ` --pypolymlp ` option, MLPs
243
- are read from ` phono3py.pmlp ` if the file exists. In this case, training data is
243
+ are read from ` polymlp.yaml ` if the file exists. In this case, training data is
244
244
no longer required, and files such as ` phono3py.yaml ` can be used as the input
245
245
structure file.
246
246
@@ -274,7 +274,7 @@ NAC parameters were read from "phono3py.yaml".
274
274
Pypolymlp is a generator of polynomial machine learning potentials.
275
275
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
276
276
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
277
- Load MLPs from "phono3py.pmlp ".
277
+ Load MLPs from "polymlp.yaml ".
278
278
------------------------------ pypolymlp end -------------------------------
279
279
Generate displacements
280
280
Displacement distance: 0.01
@@ -318,7 +318,7 @@ Random displacements are generated by specifying {ref}`--rd
318
318
<random_displacements_option>` option. To compute force constants with random
319
319
displacements, an external force constants calculator is necessary. By default,
320
320
symfc is used unless another force constants solver is explicitly specified.
321
- When running with the ` --pypolymlp ` option, MLPs are read from ` phono3py.pmlp `
321
+ When running with the ` --pypolymlp ` option, MLPs are read from ` polymlp.yaml `
322
322
if the file exists. In this case, training data is no longer required, and files
323
323
such as ` phono3py.yaml ` can be used as the input structure file.
324
324
@@ -352,7 +352,7 @@ NAC parameters were read from "phono3py.yaml".
352
352
Pypolymlp is a generator of polynomial machine learning potentials.
353
353
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
354
354
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
355
- Load MLPs from "phono3py.pmlp ".
355
+ Load MLPs from "polymlp.yaml ".
356
356
------------------------------ pypolymlp end -------------------------------
357
357
Generate random displacements
358
358
Twice of number of snapshots will be generated for plus-minus displacements.
@@ -406,10 +406,10 @@ displacements. Note that to achieve accurate force constants, the actual number
406
406
of generated supercells is twice the specified number. If ` --rd ` is omitted,
407
407
systematic displacements are introduced.
408
408
409
- Once the file ` phono3py.pmlp ` is obtained, force constants can be calculated
410
- using MLPs from ` phono3py.pmlp ` . After removing the ` fc3.hdf5 ` and ` fc2.hdf5 `
411
- files, ` phono3py-load ` will detect ` phono3py.pmlp ` and then compute the force
412
- constants by loading the MLPs from ` phono3py.pmlp ` as follows:
409
+ Once the file ` polymlp.yaml ` is obtained, force constants can be calculated
410
+ using MLPs from ` polymlp.yaml ` . After removing the ` fc3.hdf5 ` and ` fc2.hdf5 `
411
+ files, ` phono3py-load ` will detect ` polymlp.yaml ` and then compute the force
412
+ constants by loading the MLPs from ` polymlp.yaml ` as follows:
413
413
414
414
```
415
415
% phono3py-load --pypolymlp --rd 100 --amplitude 0.005 phono3py.yaml
@@ -441,7 +441,7 @@ NAC parameters were read from "phono3py.yaml".
441
441
Pypolymlp is a generator of polynomial machine learning potentials.
442
442
Please cite the paper: A. Seko, J. Appl. Phys. 133, 011101 (2023).
443
443
Pypolymlp is developed at https://github.com/sekocha/pypolymlp.
444
- Load MLPs from "phono3py.pmlp ".
444
+ Load MLPs from "polymlp.yaml ".
445
445
------------------------------ pypolymlp end -------------------------------
446
446
Generate random displacements
447
447
Twice of number of snapshots will be generated for plus-minus displacements.
@@ -542,3 +542,14 @@ the next 100 supercells) will be computed and included. With this procedure in
542
542
mind, it may be convenient to generate a sufficiently large number of supercells
543
543
with random displacements in advance, such as 1000 supercells, before starting
544
544
the LTC calculation with pypolymlp.
545
+
546
+ ## Converting ` phono3py.pmlp ` to ` polymlp.yaml `
547
+
548
+ In older versions, polynomial MLPs were stored in ` phono3py.pmlp ` . This file can
549
+ be converted to ` polymlp.yaml ` using the following Python snippet.
550
+
551
+ ``` python
552
+ from pypolymlp.mlp_dev.pypolymlp import Pypolymlp
553
+ polymlp = Pypolymlp()
554
+ polymlp.convert_to_yaml(filename_txt = " phono3py.pmlp" , filename_yaml = " polymlp.yaml”)
555
+ ```
0 commit comments