Harry Carey
Harry Carey
switching to tf.ones did not solve the issue for me on Tensorflow 2.4.1 however downgrading to 2.2 solved it.
Anyone know how?
Changing the default diam_mean to 17. in the initialisation of CPNet fixes this (and obviously breaks all other models) https://github.com/MouseLand/cellpose/blob/509ffca33737058b0b4e2e96d506514e10620eb3/cellpose/resnet_torch.py#L186-L238 So it seems like the default is not being overwritten...
Great to hear. Thank you for your response. Happy New Year!
This is a really cool idea! Do you think you could make the AIs logs public as it self improves? Either in the repo or elsewhere. I would be very...
one blocking issue at the moment is the labels appear improperly interpolated.  I've raised an issue so hopefully it will be resolved soon. https://github.com/emilyjanedennis/PRA/issues/10
Ill be able to do this soon :)
I agree this would be nice to fix Many older volumes do not have any direction cosigns set. previously this was not an issue. However since at least 4.0 opening...
I'll have more of a look later but looking through the code here it sets diam mean to 17 only if the pretrained_model_string== 'nuclei: https://github.com/MouseLand/cellpose/blob/509ffca33737058b0b4e2e96d506514e10620eb3/cellpose/models.py#L256-L259 if we make a custom...
Great idea, I will look into this after easter :)