Hello , the image folder contains both affected and un infected images is in order if i would first creat two folders one containing affected images and another one containing healthy images
@Muhire-taa, the filenames and their respective labels have been provided in CSV format so you can distinguish the classes. For the framework or approach you intend to use, it might be best to load images using dataframe or CSV methods, rather than from folders.
thanks @Professor, actually i had configured that before , but again another issue my ImageDataGenerator is only finding 52 images from the image samples and its showing that the rest are invalid images what could the cause of this issue
and wen i pass validate_file names=false in flow_from_dataframe method i get all the images but the model complains about the no such file/ directory found
Hmm, that's interesting @Muhire-taa, apologies again for my poor response speed. A friend drafted thus tutorial blog post I'll like you to check out. I'm sure your problem is solved there.
Hello @ professor , actually reffering from the above notebook am actually reffering to the above notebook am getting the error below .
Found 0 validated image filenames belonging to 0 classes.
/usr/local/lib/python3.7/dist-packages/keras_preprocessing/image/dataframe_iterator.py:282: UserWarning: Found 1619 invalid image filename(s) in x_col="Image_id". These filename(s) will be ignored.
.format(n_invalid, x_col)
And using my other working i was getting the error below
Found 42 validated image filenames belonging to 2 classes.
/usr/local/lib/python3.7/dist-packages/keras_preprocessing/image/dataframe_iterator.py:282: UserWarning: Found 1253 invalid image filename(s) in x_col="Image_id". These filename(s) will be ignored.
.format(n_invalid, x_col)
and when i set the validate_filenames=False in the flow_from_dataframe method i get the below
Found 324 non-validated image filenames belonging to 2 classes( but this does not work out since the images are not_validated)
How can i go about the above issues
Is it because that am working in google colab
@Muhire-taa, the filenames and their respective labels have been provided in CSV format so you can distinguish the classes. For the framework or approach you intend to use, it might be best to load images using dataframe or CSV methods, rather than from folders.
but when i run ImageDataGenerator ,am getting found 1000 images belonging to 1 class,but thy should be two classes
Apologies for the late response @muhire-taa, but yeah, you should have two classes. Perherps there's a bug in your code. You'll need to troubleshoot.
thanks @Professor, actually i had configured that before , but again another issue my ImageDataGenerator is only finding 52 images from the image samples and its showing that the rest are invalid images what could the cause of this issue
Thanks
and wen i pass validate_file names=false in flow_from_dataframe method i get all the images but the model complains about the no such file/ directory found
Hmm, that's interesting @Muhire-taa, apologies again for my poor response speed. A friend drafted thus tutorial blog post I'll like you to check out. I'm sure your problem is solved there.
https://medium.com/@babatundeoreoluwa35/image-classification-using-transfer-learning-crop-disease-classification-dc4dedab17cb
I'm happy to help if you still don't get it resolved.
Hello @ professor , actually reffering from the above notebook am actually reffering to the above notebook am getting the error below .