[BUG] PT: passing >f4 array to DeepPot triggers ValueError

[njzjz]

Bug summary

It seems to me that passing a >f4 array to DeepPot will trigger ValueError: given numpy array has byte order different from the native byte order. Conversion between byte orders is currently not supported.

>f4 may be the original data type stored in the NetCDF files.

DeePMD-kit Version

5111e9b

Backend and its version

PyTorch 2.3.1

How did you download the software?

Built from source

Input Files, Running Commands, Error Log, etc.

Traceback (most recent call last):
  File "/home/jz748/codes/deepmd-kit/examples/water/se_e2_a/1.py", line 8, in <module>
    e, f, v = dp.eval(coord, cell, atype)
  File "/home/jz748/codes/deepmd-kit/deepmd/infer/deep_pot.py", line 201, in eval
    results = self.deep_eval.eval(
  File "/home/jz748/codes/deepmd-kit/deepmd/pt/infer/deep_eval.py", line 268, in eval
    out = self._eval_func(self._eval_model, numb_test, natoms)(
  File "/home/jz748/codes/deepmd-kit/deepmd/pt/infer/deep_eval.py", line 340, in eval_func
    return self.auto_batch_size.execute_all(
  File "/home/jz748/codes/deepmd-kit/deepmd/utils/batch_size.py", line 203, in execute_all
    n_batch, result = self.execute(execute_with_batch_size, index, natoms)
  File "/home/jz748/codes/deepmd-kit/deepmd/utils/batch_size.py", line 117, in execute
    raise e
  File "/home/jz748/codes/deepmd-kit/deepmd/utils/batch_size.py", line 114, in execute
    n_batch, result = callable(max(batch_nframes, 1), start_index)
  File "/home/jz748/codes/deepmd-kit/deepmd/utils/batch_size.py", line 180, in execute_with_batch_size
    return (end_index - start_index), callable(
  File "/home/jz748/codes/deepmd-kit/deepmd/pt/infer/deep_eval.py", line 383, in _eval_model
    coord_input = torch.tensor(
ValueError: given numpy array has byte order different from the native byte order. Conversion between byte orders is currently not supported.

Steps to Reproduce

from deepmd.infer import DeepPot
import numpy as np

dp = DeepPot("frozen_model.pth")
coord = np.array([[1, 0, 0], [0, 0, 1.5], [1, 0, 3]], dtype=">f4").reshape([1, -1])                                                                                                                                                                                                                                          
cell = np.diag(10 * np.ones(3)).reshape([1, -1])
atype = [1, 0, 1]
e, f, v = dp.eval(coord, cell, atype)

Further Information, Files, and Links

No response