[BUG] Cellpose 3.0.7 does not remember that turning off auto-adjust saturation during image training
Closed this issue · 1 comments
Describe the bug
Cellpose 3.0.7 does not remember that turning off auto-adjust saturation during image training.
To Reproduce
Steps to reproduce the behavior:
- deactivate auto-adjust saturation
- perform training on image
- newly loaded image has autoadjusted saturation on again
also happens without training if any type of model is used to segment image
Logs:
(cellpose) C:\Users\leone>Cellpose
2024-04-18 15:25:55,577 [INFO] WRITING LOG OUTPUT TO C:\Users\leone.cellpose\run.log
2024-04-18 15:25:55,577 [INFO]
cellpose version: 3.0.7
platform: win32
python version: 3.8.19
torch version: 2.2.2+cu118
2024-04-18 15:25:55,792 [INFO] ** TORCH CUDA version installed and working. **
GUI_INFO: loading image: C:/Users/leone/Documents/SingleCell Local/WT mEOS Comparison/M1 Choosen/Training 18.4.24 Flowcell WT mEOS + WT vom 20.10.23/mEOS6-94.tif
GUI_INFO: selected model new Training flowcell+wt old, loading now
2024-04-18 15:28:03,801 [INFO] >> new Training flowcell+wt old << model set to be used
2024-04-18 15:28:03,801 [INFO] ** TORCH CUDA version installed and working. **
2024-04-18 15:28:03,801 [INFO] >>>> using GPU
2024-04-18 15:28:03,886 [INFO] >>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
2024-04-18 15:28:03,886 [INFO] >>>> model diam_labels = 21.940 (mean diameter of training ROIs)
GUI_INFO: diameter set to 21.94 (but can be changed)
2024-04-18 15:28:05,757 [INFO] >>>> loading model C:\Users\leone.cellpose\models\new Training flowcell+wt old
2024-04-18 15:28:05,776 [INFO] ** TORCH CUDA version installed and working. **
2024-04-18 15:28:05,776 [INFO] >>>> using GPU
2024-04-18 15:28:05,857 [INFO] >>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
2024-04-18 15:28:05,857 [INFO] >>>> model diam_labels = 21.940 (mean diameter of training ROIs)
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
2024-04-18 15:28:07,621 [INFO] 7 cells found with model in 1.863 sec
GUI_INFO: 7 masks found
GUI_INFO: creating cellcolors and drawing masks
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
[0, 255.0]
2024-04-18 15:28:16,551 [INFO] >>>> loading model C:\Users\leone.cellpose\models\new Training flowcell+wt old
2024-04-18 15:28:16,551 [INFO] ** TORCH CUDA version installed and working. **
2024-04-18 15:28:16,551 [INFO] >>>> using GPU
2024-04-18 15:28:16,651 [INFO] >>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
2024-04-18 15:28:16,651 [INFO] >>>> model diam_labels = 21.940 (mean diameter of training ROIs)
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
2024-04-18 15:28:17,065 [INFO] 7 cells found with model in 0.513 sec
GUI_INFO: 7 masks found
GUI_INFO: creating cellcolors and drawing masks
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
[0, 255.0]
2024-04-18 15:28:22,510 [INFO] ** TORCH CUDA version installed and working. **
2024-04-18 15:28:22,510 [INFO] >>>> using GPU
2024-04-18 15:28:22,517 [INFO] >> cyto3 << model set to be used
2024-04-18 15:28:22,595 [INFO] >>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
2024-04-18 15:28:22,595 [INFO] channels set to [0, 0]
2024-04-18 15:28:22,611 [INFO] ~~~ FINDING MASKS ~~~
2024-04-18 15:28:23,012 [INFO] >>>> TOTAL TIME 0.42 sec
2024-04-18 15:28:23,018 [INFO] 7 cells found with model in 0.508 sec
GUI_INFO: 7 masks found
GUI_INFO: creating cellcolors and drawing masks
{'lowhigh': None, 'percentile': [1.0, 99.0], 'normalize': True, 'norm3D': False, 'sharpen_radius': 0, 'smooth_radius': 0, 'tile_norm_blocksize': 0, 'tile_norm_smooth3D': 1, 'invert': False}
[0, 255.0]
thanks for reporting this, it should be fixed now: pip install git+https://github.com/mouseland/cellpose.git