No positive labels for row {i}

[tannaz-k]

Hi I download DiabeticRetinopathyDetection code, run setup.h, download kaggle dataset as you said and then run the preprocess.py.All done successfully. I also rename the configuration.py and edit the parameters. But when I run train_model.py, the following error occured:

Traceback (most recent call last): File "Desktop/DiabeticRetinopathyDetection-master2/src/train_model.py", line 43, in <module> pin_memory=True) File "/home/linux/Desktop/DiabeticRetinopathyDetection-master2/src/data_loading.py", line 181, in get_train_valid_loader train_idx = label_balancer.rebalance_categorical_train_idxs_evenly() File "/home/linux/Desktop/DiabeticRetinopathyDetection-master2/src/data_loading.py", line 99, in rebalance_categorical_train_idxs_evenly row_pos, row_n = self._get_row_counts(num_classes) File "/home/linux/Desktop/DiabeticRetinopathyDetection-master2/src/data_loading.py", line 67, in _get_row_counts raise ValueError(f"No positive labels for row {i}.") ValueError: No positive labels for row 3. I would appreciate your help :( , Thanks in advanced

[saurabheights]

Hi, downloading the whole dataset is not possible for me right now. Could you upload your configuration file? I can try running it on sample dataset file and see if the issue persists.

[saurabheights]

P.S. In case you are running on sample data of 10 images, make sure out of those 10 images, there is at least one image for each label 0 to 4.

[tannaz-k]

Hi I run the code on the sample data that you've uploaded. There's no image for class 3, I suppose. Here's my config file:

configuration.zip And here's my sample data:

[sample.zip] (https://github.com/saurabheights/DiabeticRetinopathyDetection/files/2402007/sample.zip)

[tannaz-k]

P. S. I've added one image for class 3 in the train dataset and this error occured:

Traceback (most recent call last): File "Desktop/DiabeticRetinopathyDetection-master2/src/train_model.py", line 52, in <module> model = model_params['model'](**model_params['model_kwargs'], default_lr=optimizer_params['lr']) TypeError: __init__() got an unexpected keyword argument 'default_lr'

[saurabheights]

Hi, I will look into it as soon as possible but currently tied in office work. If you want, you can try changing the learning params and rebalancing strategy(set to None) till I reply.

[tannaz-k]

Hi, thank you for your prompt reply. I've done what you said, but there's still an error. I changed the learning params and set rebalancing strategy to None as u said. Am I doing sth wrong? Or it's a bug in the code?

Traceback (most recent call last): File "Desktop/DiabeticRetinopathyDetection-master2/src/train_model.py", line 52, in <module> model = model_params['model'](**model_params['model_kwargs'], default_lr=optimizer_params['lr']) TypeError: __init__() got an unexpected keyword argument 'default_lr'

[omsh]

Hi, Use the transfer learning configuration file (here). You can also change the 'model' to DRNet in your configuration.

[tannaz-k]

Hi, Thanks a lot, I will try it and tell you the result.

[tannaz-k]

Hi, I changed my config file to the file that you've told and run it on kaggle dataset. I also change batch size to 10. Then I execute sudo command sudo python3 Desktop/DiabeticRetinopathyDetection-master2/src/train_model.py.and the training has finished till epoch 50. At the end of epoch 50, this error has occured:

Traceback (most recent call last):
  File "Desktop/DiabeticRetinopathyDetection-master2/src/train_model.py", line 95, in <module>
    predictions, image_names = model_trainer.predict_on_test()
  File "/home/linux/Desktop/DiabeticRetinopathyDetection-master2/src/trainer.py", line 178, in predict_on_test
    bar = progressbar.ProgressBar(max_value=len(self.test_dataset_loader))
TypeError: __init__() got an unexpected keyword argument 'max_value'

[saurabheights]

Dear, please try to google these issues first. It will save both your and my time. Progressbar is just to display how much of work is done, the same way apt-get shows a horizontal bar. The second result of the error shows a possible solution:- https://github.com/ikostrikov/TensorFlow-VAE-GAN-DRAW/issues/8 - just change max_value maxval.

You can also check the code where error happened and quick fix it.

Second, please create a new issue page for different problem faced.

P.S. - I will fix these issue in the repo, as soon as I resolve my current tasks.

[mrzhangzizhen123]

Hi, Use the transfer learning configuration file (here). You can also change the 'model' to DRNet in your configuration.

Could you explain the configuration details in detail?I made a mistake according to your configuration.

[mrzhangzizhen123]

@omsh Could you explain the configuration details in detail?I made a mistake according to your configuration.

[mrzhangzizhen123]

@omsh You can provide your bug-free code

[mrzhangzizhen123]

@tannaz-k can you provide your bug-free code？