Seeking help in implementing cyto-dl/labelfree for 2D images #314
Unanswered
KartikeyKansal1
asked this question in
Q&A
Replies: 1 comment 5 replies
-
|
Hi @KartikeyKansal1! Thanks for your interest in our tool and feedback on the docs. We set This should load your 2d images as a CYX dimension array, which the rest of the transforms expect. You don't need to create a zstack of your 2d images, just specifying paths in the |
Beta Was this translation helpful? Give feedback.
5 replies
-
|
Hey @benjijamorris, thank you for your swift response. I am still running into errors, unfortunately. |
Beta Was this translation helpful? Give feedback.
-
|
Ah, sorry! I forgot that the |
Beta Was this translation helpful? Give feedback.
-
|
Hi @benjijamorris , thank you for getting back, this worked perfectly :) I'm taking my This is how my experiment/labelfree.yaml looks like: path in config/data/labelfree.yaml refers to a csv file with only brightfield images. I've changed the save_dir in config/model file itself |
Beta Was this translation helpful? Give feedback.
-
|
Hi @KartikeyKansal1 sorry for the late reply! Adding a column called |
Beta Was this translation helpful? Give feedback.
-
|
Hi @benjijamorris , thank you for getting back and helping me with the implementation of cyto-dl. Sorry for my late reply: this resolved all the errors! Though I still don't see any generated images saved in my save_dir directory. I see that it created 4 sub-directories('predict_images', 'test_images', 'train_images', 'val_images') after I ran the code but all of them are empty. I am following this documentation so far: https://github.com/AllenCellModeling/cyto-dl/blob/main/docs/using_examples.rst Update: I removed one image deliberately from 'images to predict on dataset' and still had the path for it in the predict.csv file, the code didn't break, in contrast to during training when I remove one file from the training dataset while its path is still present in the train.csv, the code breaks; which leads to me to believe that it is still not able to read the files to run the prediction on. |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi, I have many 2D images in tif format for the model to train on. I've changed my spatial_dims=2, patch_shape=[64,64] in config files, and updated the data paths in the CSV file in data/labelfree. Specified the channel:1 instead of 5/6 used in the example 3D image and changed the z=1 instead of z=38. Not sure why z was defaulted to 38.
In particular, I couldn't find any documentation if I need to create a z-stack of my 2D images? or list them sequentially in the CSV files with data path? I'm currently trying to run train.py. The code works fine with the example data.
I've pasted the complete error message below:
Beta Was this translation helpful? Give feedback.
All reactions