[BUG]NET ERROR: tensors used as indices must be long, byte or bool tensors

[sun1000yao]

NET ERROR: tensors used as indices must be long, byte or bool tensors

Hi, When I was using cellpose Zstack to run cyto3 with GPU, this error popped up. Same issue appeared when uncheck the 'use GPU'. Thanks in advance for suggestions and helps.

Run log
2024-12-19 23:37:03,114 [INFO] ** TORCH CUDA version installed and working. **
2024-12-19 23:37:03,114 [INFO] >>>> using GPU (CUDA)
2024-12-19 23:37:03,115 [INFO] >> cyto3 << model set to be used
2024-12-19 23:37:03,247 [INFO] >>>> loading model C:\Users\Bagnall_Lab.cellpose\models\cyto3
2024-12-19 23:37:06,428 [INFO] >>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': True, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
2024-12-19 23:37:06,440 [INFO] channels set to [0, 0]
2024-12-19 23:37:06,441 [INFO] ~~~ FINDING MASKS ~~~
2024-12-19 23:37:06,441 [INFO] multi-stack tiff read in as having 43 planes 1 channels
2024-12-19 23:37:06,580 [INFO] running YX: 43 planes of size (399, 911)
2024-12-19 23:37:06,581 [INFO] 0%| | 0/43 [00:00<?, ?it/s]
2024-12-19 23:37:22,244 [INFO] 100%|##########| 43/43 [00:15<00:00, 2.75it/s]
2024-12-19 23:37:22,330 [INFO] running ZY: 399 planes of size (43, 911)
2024-12-19 23:37:22,330 [INFO] 0%| | 0/399 [00:00<?, ?it/s]
2024-12-19 23:37:29,731 [INFO] 100%|##########| 399/399 [00:07<00:00, 53.91it/s]
2024-12-19 23:37:29,804 [INFO] running ZX: 911 planes of size (43, 399)
2024-12-19 23:37:29,804 [INFO] 0%| | 0/456 [00:00<?, ?it/s]
2024-12-19 23:37:39,267 [INFO] 100%|##########| 456/456 [00:09<00:00, 48.19it/s]
2024-12-19 23:37:39,627 [INFO] network run in 33.05s
NET ERROR: tensors used as indices must be long, byte or bool tensors

Package version
cellpose 3.1.0 pypi_0 pypi
cudatoolkit 11.3.1 h59b6b97_2
numpy 1.26.3 pypi_0 pypi
python 3.10.16 h4607a30_1
pytorch 1.12.0 py3.10_cuda11.3_cudnn8_0 pytorch
pyqt6 6.8.0 pypi_0 pypi
pyqtgraph 0.13.7 pypi_0 pypi
numba 0.60.0 pypi_0 pypi