Argument data does not have the correct dtype

[robinbenjamineberle]

upon trying the example dataset and running

python3 convert_particle_stacks_example.py

the output looks like below. I tried to play around with the nufft.py and forcing torch.complex128. unsuccessfully.

python3 convert_particle_stacks_example.py
pixel_size: [1.346 1.346]
Warning: Invalid precision value, using default
Warning: Invalid atom shape value, using default
Parameters:
n_voxels: [132 132 132]
voxel_size: [1.346 1.346 1.346]
box_size: [177.672 177.672 177.672]
radius_max: 33.0
dist_radii: 0.25
n_inplanes: 256
precision: Precision.DEFAULT
viewing_distance: 0.6366197723675814
atom_radii: 3.0
atom_selection: name CA
use_protein_residue_model: True
atom_shape: AtomShape.DEFAULT
Warning: CTF amplitude contrast not found, using default value 0.1
Warning: CTF phase shift not found, using default value 0.0
Warning: CTF B-factor not found, using default value 0.0
Warning: CTF scale factor not found, using default value 1.0
Physical images shape: torch.Size([483, 132, 132])
Traceback (most recent call last):
File "/home/CryoLike/example/convert_particle_stacks_example.py", line 9, in
convert_particle_stacks_from_star_files(
File "/home/venv/lib/python3.11/site-packages/cryolike/convert_particle_stacks/particle_stacks_conversion.py", line 197, in convert_particle_stacks_from_star_files
_do_image_normalization(im_batch, polar_grid, params.precision)
File "/home/venv/lib/python3.11/site-packages/cryolike/convert_particle_stacks/particle_stacks_conversion.py", line 59, in _do_image_normalization
im.transform_to_fourier(polar_grid=polar_grid, precision=precision, use_cuda=True)
File "/home/venv/lib/python3.11/site-packages/cryolike/stacks/image.py", line 371, in transform_to_fourier
self.images_fourier = cartesian_phys_to_fourier_polar(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/venv/lib/python3.11/site-packages/cryolike/microscopy/nufft.py", line 205, in cartesian_phys_to_fourier_polar
image_polar[i,:] = nufft2d2_cpu(k1, k2, images_phys[i], eps = eps, isign = isign)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/venv/lib/python3.11/site-packages/finufft/_interfaces.py", line 825, in nufft2d2
return invoke_guru(2,2,x,y,None,out,None,None,None,f,isign,eps,None,**kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/venv/lib/python3.11/site-packages/finufft/_interfaces.py", line 561, in invoke_guru
out = plan.execute(f,c)
^^^^^^^^^^^^^^^^^
File "/home/venv/lib/python3.11/site-packages/finufft/_interfaces.py", line 249, in execute
_data = _ensure_array_type(data, "data", self.dtype)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/venv/lib/python3.11/site-packages/finufft/_interfaces.py", line 317, in _ensure_array_type
raise TypeError(f"Argument {name} does not have the correct dtype: {x.dtype} was given, but {dtype} was expected.")
TypeError: Argument data does not have the correct dtype: complex64 was given, but complex128 was expected.