Numpy version issue

[samholtzen]

Tried using cellSAM by running the following:

from cellSAM import segment_cellular_image, get_model
import numpy as np
import torch
from scipy.ndimage import binary_dilation
import matplotlib.pyplot as plt
import matplotlib.patches as patches
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

img = image[0,...,2]
print(img.dtype)
mask, embedding, bounding_boxes = segment_cellular_image(np.array(img), device=str(device))

plt.imshow(mask)

with the most recent version of Numpy and it gave me the error:

TypeError: expected np.ndarray (got numpy.ndarray)

When I rolled back the version to numpy==1.23.5, I was able to run the code block.

[rossbar]

Thanks @samholtzen - this too is a consequence of the torchvision pin, which also then requires numpy<2. This will be solved by #17 - thanks for reporting!