sneddy / pneumothorax-segmentation Goto Github PK

can you help to provide the processed dataset, thk

Hi @sneddy
I had a look at the issue #6 , but the same Error occurs to me in-between training. I have attached a screenshot for your convenience.

I have also checked for that specific file. The image and its segmentation mask are present in the directory. There seems to be no error when I open the same through OpenCV from another python file.

Hi @sneddy, thanks for sharing your great work!
I'm having an issue running your code after downloading it.
I'm getting this error, how do I resolve it?

Train.py: error: the following arguments are required: train_cfg

I can't trace the error, can anybody tell me what do I have to change to make this run?

I have tried your method to do joint transforms but failed. The target did not changed at all.
need your help, thank you.

I can't reproduce your results, can you provide your pretrain code or tell me how to pretrain to get better grade?

Thanks for your code. I want to use your code to test my experimental results in my paper. What do I need to cite in my reference?

Hi @sneddy
I am trying to implement your model. It works fine until this line
python Train.py experiments/albunet_valid/train_config_part0.yaml.
On running this it gives an error as Nonetype object has no attribute shape. I think that the dataset is not being served. I have made changes to the path in Pneumadataset.py as well as the DATA_DIRECTORY in train_config_part0.yaml to my path and run the Train.py file. But it still doesn't work. Am I missing out on anything? It might be a silly error but any help would be appreciated.
Attaching the screenshot of the error

validation during experiment 0 exhausts whole ram memory. Could you please have a look at that

Is there any specific pattern in the identifying the images for different folds ?
How was this file train_folds_5.csv generated ?

Hi, thanks for the great work.
I tried reproducing your performance following the instruction. (Albunet_valid_1024)

After part 0 training, the best performance for 5-folds looks just 0.78 ( I checked experiment/almunet_valid/checkpoint/summary.csv), which is lower than I think (0.85),
could you let me know how to get 0.85 of dice score on validation set?