Skip to content

1st place solution of Kaggle Open Problems - Multimodal Single-Cell Integration

License

Notifications You must be signed in to change notification settings

shu65/open-problems-multimodal

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

open-problems-multimodal

1st place solution of Kaggle Open Problems - Multimodal Single-Cell Integration

Preparation

Install the solution code.

pip3 install -e .

In addtion, download the following data

  1. Open Problems - Multimodal Single-Cell Integration data set from Kaggle
  2. tab separated hgnc_complete_set file from https://www.genenames.org/download/archive/
  3. Reactome Pathways Gene Set from https://reactome.org/download-data

Compress data and make addtitional data

compress kaggle dataset and make addtional data to use in training

export DATA_DIR=/path/to/kaggle/dataset/Directory
python3 script/make_compressed_dataset.py --data_dir ${DATA_DIR}
python3 script/make_additional_files.py --data_dir ${DATA_DIR}
python3 script/make_cite_input_mask.py --data_dir ${DATA_DIR} --hgnc_complete_set_path /path/to/hgnc_complete_set --reactome_pathways_path /path/to/reactome_pathways

Training

Multi

python3 scripts/train_mode.py --data_dir ${DATA_DIR} --task_type multi 

Cite

python3 scripts/train_mode.py --data_dir ${DATA_DIR} --task_type cite 

About

1st place solution of Kaggle Open Problems - Multimodal Single-Cell Integration

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages