FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv

[Rain-0908]

When I run the command "python bin/main.py --base_path EMPaCT --config_path configs/i2i_config.yml --src_marker HE --dst_marker NKX3 --is_train True",it gives an error.-------------------FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv' ----------------------------------------------------------------------------------------
I am confused that EMPaCT dataset is downloaded from the https://zenodo.org/records/10066853.Exactly,there is no file named HE_splits.csv . So what the problem is?Can you give me some tips?

[Yifever20002]

same problem, any update?

[ZhouChen-start]

Yes, I have the same problem. The link provided by the author seems to be the raw data, and no data processing code is provided.

[MeghdadSabouri]

same problem

[ZhouChen-start]

When I run the command "python bin/main.py --base_path EMPaCT --config_path configs/i2i_config.yml --src_marker HE --dst_marker NKX3 --is_train True",it gives an error.-------------------FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv' ---------------------------------------------------------------------------------------- I am confused that EMPaCT dataset is downloaded from the https://zenodo.org/records/10066853.Exactly,there is no file named HE_splits.csv . So what the problem is?Can you give me some tips?

Hi, have you found a solution?

[ZhouChen-start]

same problem, any update?

Hi, have you found a solution?

[Rain-0908]

same problem, any update?

Hi, have you found a solu

[Rain-0908]

When I run the command "python bin/main.py --base_path EMPaCT --config_path configs/i2i_config.yml --src_marker HE --dst_marker NKX3 --is_train True",it gives an error.-------------------FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv' ---------------------------------------------------------------------------------------- I am confused that EMPaCT dataset is downloaded from the https://zenodo.org/records/10066853.Exactly,there is no file named HE_splits.csv . So what the problem is?Can you give me some tips?

Hi, have you found a solution?
I think the data set provided by the data set website lacks a text description file (csv), which has not been resolved yet. If you solve it, please share it.

[ZhouChen-start]

When I run the command "python bin/main.py --base_path EMPaCT --config_path configs/i2i_config.yml --src_marker HE --dst_marker NKX3 --is_train True",it gives an error.-------------------FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv' ---------------------------------------------------------------------------------------- I am confused that EMPaCT dataset is downloaded from the https://zenodo.org/records/10066853.Exactly,there is no file named HE_splits.csv . So what the problem is?Can you give me some tips?

Hi, have you found a solution?
I think the data set provided by the data set website lacks a text description file (csv), which has not been resolved yet. If you solve it, please share it.

I have debugged the entire project and found that if the author does not provide complete dataset information, no one will be able to reproduce the project. I can give you the following information: 1. The .csv file contains two columns, trian_cores and trian_cores, which are the image file names without the png suffix. 2. There is a folder named NKX3 in the file downloaded from https://zenodo.org/records/10066853, but NKX is used in the code. You can pay attention to it. 3. In train.py, model = create_model(args, device), the train.model.model_name parameter is missing. You can add it through args['train.model.model_name'] = 'XXX'. XXX should be a subclass of BaseModel. 4. The dataset that needs to be loaded in the project should be cut into 512x512 patches. Since the author did not provide a script for segmentation, I tried to write a script myself and rename it according to the information found by debugging, such as A1_TMA_15_02_IB_HE_y_0_x_0_patch_.png. 5. The last point is also the one that I think cannot be reproduced. The .npz file corresponding to the patch is missing. There is a very important information in this file. Through the two lines of code bbox = np.load(os.path.join(bbox_info_path, basename + '.npz'))
bbox = bbox['bbox'], we can know that the 'bbox' information is missing. The content should be y0, x0, y1, x1, roiclass (0: pos roi, 1: neg roi).
The above is the information I know so far. If you have more useful information, please share it with me. Thank you very much!

[Yifever20002]

When I run the command "python bin/main.py --base_path EMPaCT --config_path configs/i2i_config.yml --src_marker HE --dst_marker NKX3 --is_train True",it gives an error.-------------------FileNotFoundError: [Errno 2] No such file or directory: 'EMPaCT/data_splits/i2i/HE_splits.csv' ---------------------------------------------------------------------------------------- I am confused that EMPaCT dataset is downloaded from the https://zenodo.org/records/10066853.Exactly,there is no file named HE_splits.csv . So what the problem is?Can you give me some tips?

Hi, have you found a solution?
I think the data set provided by the data set website lacks a text description file (csv), which has not been resolved yet. If you solve it, please share it.

I have debugged the entire project and found that if the author does not provide complete dataset information, no one will be able to reproduce the project. I can give you the following information: 1. The .csv file contains two columns, trian_cores and trian_cores, which are the image file names without the png suffix. 2. There is a folder named NKX3 in the file downloaded from https://zenodo.org/records/10066853, but NKX is used in the code. You can pay attention to it. 3. In train.py, model = create_model(args, device), the train.model.model_name parameter is missing. You can add it through args['train.model.model_name'] = 'XXX'. XXX should be a subclass of BaseModel. 4. The dataset that needs to be loaded in the project should be cut into 512x512 patches. Since the author did not provide a script for segmentation, I tried to write a script myself and rename it according to the information found by debugging, such as A1_TMA_15_02_IB_HE_y_0_x_0_patch_.png. 5. The last point is also the one that I think cannot be reproduced. The .npz file corresponding to the patch is missing. There is a very important information in this file. Through the two lines of code bbox = np.load(os.path.join(bbox_info_path, basename + '.npz')) bbox = bbox['bbox'], we can know that the 'bbox' information is missing. The content should be y0, x0, y1, x1, roiclass (0: pos roi, 1: neg roi). The above is the information I know so far. If you have more useful information, please share it with me. Thank you very much!

Thank you! This is really helpful! Actually I'm also facing the same problem in your step 5, and seems there's nothing we can do except for acquiring the labeled roiclass. By the way, @rapsoman could you pay little attention to this? It's somehow more frustrating to have a half-open code than a closed-source one.

[rapsoman]

Dear all, apologies for the issue, which we are looking into and will take care of soon. I kindly ask you for some patience while we do. Many thanks!

[MayukhmalaJana]

Dear @rapsoman,

Thank you for your efforts, I am looking forward to the updated code as I have been facing several issues regarding this. Is it possible to train the model by providing input of HE images and IHC images?

[pushpak-pati]

Thank you for your queries.
The repository is now updated with relevant preprocessing code at https://github.com/AI4SCR/VirtualMultiplexer/tree/master/preprocess.

The additional relevant data (data splits, tissue masks, bounding boxes) are now available at https://figshare.com/projects/VirtualMultiplexer/230498