search

xianglin226 / scMDC

Single Cell Multi-omics deep clustering
Apache License 2.0
24 stars 4 forks source link

run_scMDC.py output folder is empty? #13

Open kabilanbio opened 5 months ago

kabilanbio commented 5 months ago

Your work is very useful to single-cell multiomics anlaysis. I have tried your model to run in Google colab L4 GPU with the example BMNC CITE-seq dataset. I have completely run the model for 3 times with 400 epochs for pretraining and around 250 epochs for the clustering stage. But every time I am getting only the output of 'AE_weights_bmnc' only. In the output folder named 'citeseq_bmnc' is always empty. Is there any way to solve this problem?

Screenshot 2024-07-02 014542

xianglin226 commented 4 months ago

Thank you for reaching out. Did you set --embedding_file and --prediction_file as true?

kabilanbio commented 4 months ago

Thank you for your answer. How do you mention --embedding_file and --prediction_file as true in the below script? Because I have tried and got the error message.

!source activate scMDC && python -u /content/drive/MyDrive/scMDC/run_scMDC.py --n_clusters 27 --ae_weight_file AE_weights_bmnc.pth.tar --data_file $f --save_dir citeseq_bmnc --embedding_file --prediction_file --filter1 --filter2

xianglin226 commented 2 months ago

Sorry for the late reply. What error message did you get from here?

kabilanbio commented 2 months ago

I am also sorry for the late reply to your question. I ran the above-mentioned Python script with your example dataset in Google Colab. However, the script ran successfully, and clustering was done completely, I didn't find any output files in the output folder. If there is any error in the script, kindly correct it. I don't know exactly how to mention the --embedding_file and --prediction file in the script. Please help me to solve this