mahmoodlab / CLAM

Data-efficient and weakly supervised computational pathology on whole slide images - Nature Biomedical Engineering
http://clam.mahmoodlab.org
GNU General Public License v3.0
975 stars 329 forks source link

Datasets in configuration file and GPU Training Step #245

Open rohith2011 opened 2 months ago

rohith2011 commented 2 months ago

I have used the .tif files (I have manually converted the each frame of the .scn format and saved them within the range of 1xe8 constrain of the data) then I have created the patches using the preset of bwh_biopsy.csv.

Then I have followed the steps to extract the features form this data. Then I have worked and set up the DATA_ROOT_DIR for both the classes ("task_1_tumor_vs_normal").

When I setup the data_dir file directory in the if args.task == 'task_1_tumor_vs_normal': args.n_classes=2 dataset = Generic_MIL_Dataset(csv_path = 'dataset_csv/tumor_vs_normal_dummy_clean.csv', data_dir= os.path.join(args.data_root_dir, 'tumor_vs_normal_feat_resnet'), shuffle = False, seed = args.seed, print_info = True, label_dict = {'normal_tissue':0, 'tumor_tissue':1}, label_col = 'label', ignore=[])

the directory should be that of the DATA_ROOT_DIR/ ├──DATASET_1_DATA_DIR/ ├── h5_files ├── slide_1.h5 ├── slide_2.h5 └── ... └── pt_files ├── slide_1.pt ├── slide_2.pt └── ... ├──DATASET_2_DATA_DIR/ ├── h5_files ├── slide_a.h5 ├── slide_b.h5 └── ... └── pt_files ├── slide_a.pt ├── slide_b.pt └── ... └──DATASET_3_DATA_DIR/ ├── h5_files ├── slide_i.h5 ├── slide_ii.h5 └── ... └── pt_files ├── slide_i.pt ├── slide_ii.pt └── ... └── ...

I put this as the directory for the data_dir arg in the command above.

Below is the error I am getting.

FileNotFoundError: [Errno 2] No such file or directory: 'DATA_ROOT_DIR/FEATURES_DIRECTORY/pt_files/O3594401.scn - Series 2.ome.pt'

How can this be the error solve this as the directory should be like "DATA_ROOT_DIR" insted of "DATA_ROOT_DIR/FEATURES_DIRECTORY/pt_files/"

The PC configuration:-

OS-Linux GPU: A100 Nvidia