[BUG] Different Results in run_Cellpose-SAM.ipynb and https://huggingface.co/spaces/mouseland/cellpose

[akoslova]

When uploading one microscopic image on https://huggingface.co/spaces/mouseland/cellpose, the segmentation works very well, but when using the notebook run_Cellpose-SAM.ipynb, the segmentation is much worse.

Web Application:

https://huggingface.co/spaces/mouseland/cellpose

Notebook run_Cellpose-SAM.ipynb:

Channels: 0, 1, 2
flow_threshold = 0.4
cellprob_threshold = 0.0
tile_norm_blocksize = 0

I am not sure if this is a bug or if I am missing something crucial. Can you please help me to reproduce the same result in the jupyter notebook as in the manual upload? Thank you so much in advance!