External Email - Use Caution
Dear FreeSurfer experts,
I have recently upgraded to the latest stable version, 7.4.1, in order to try out the new features of SynthSeg and SynthStrip. However, I have encountered some errors. My laptop's chip is M2. Could someone please provide some guidance?
Initially, I was able to run mri_synthstrip, but not the mri_synthreg. I got the error message as:
(base) MacBook-Pro-M2:$ mri_synthseg -i ./mci_test/mri/orig.mgz -o ./mci_test/ai_test --threads 8
Traceback (most recent call last):
File "/Applications/freesurfer/7.4.1/python/scripts/mri_synthseg", line 10, in <module>
import surfa as sf
File "/Applications/freesurfer/7.4.1/python/lib/python3.8/site-packages/surfa/__init__.py", line 19, in <module>
from .image import Volume
File "/Applications/freesurfer/7.4.1/python/lib/python3.8/site-packages/surfa/image/__init__.py", line 1, in <module>
from .framed import FramedImage
File "/Applications/freesurfer/7.4.1/python/lib/python3.8/site-packages/surfa/image/framed.py", line 24, in <module>
from surfa.image.interp import interpolate
File "surfa/image/interp.pyx", line 1, in init surfa.image.interp
ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject
Enlightened by #msg74713 and #msg73091, I upgraded the numpy version from 1.19 to 1.20.3, and also installed a python specific for M1/M2.
$ sudo /Applications/freesurfer/dev/python/bin/python3 -m pip install numpy==1.20.3
(I couldn't use $ sudo fspython -m pip install, because it said my freesurfer is not properly sources, which is weird since I am sure it was sourced in a right way.
But now both mri_synthseg and mri_synthstrip can't run.
Now if I enter the commands $ fspython -m pip freeze, the outputs are:
...
neurite==0.2
nibabel==5.1.0
numpy==1.20.3
oauthlib==3.2.2
opencv-python==4.7.0.72
...
tensorboard==2.11.2
tensorboard-data-server==0.6.1
tensorboard-plugin-wit==1.8.1
tensorflow @ file:///Users/mu/Desktop/tensorflow-2.4.1-py3-none-any.whl
tensorflow-estimator==2.4.0
tensorflow-io-gcs-filesystem==0.32.0
Best,
Mu