Skip to content

Latest commit

 

History

History
36 lines (17 loc) · 879 Bytes

manual.md

File metadata and controls

36 lines (17 loc) · 879 Bytes

Gene-Disease Associations Prediction Task Manual

Data Format

Pretraining data in the directory ~/dpa_pretrain/src/pretrain/data/pretrain/disgnet.csv

Finetuning data Pretrain-Finetune in the directory ~/dpa_pretrain/src/data/downstream/

Finetuning data TDC is API, you can run it directly.

Model

You can check the structure of the model.

Pretraining model in the directory ~/dpa_pretrain/src/pretrain

Finetuning model in the directory ~/dpa_pretrain/src/finetune

Parameters

You can experiment with different parameters to find the best performance.

You adjust the Pretraining/Finetuning parameters in this script file.

In the directory ~/dpa_pretrain/src/scripts

You adjust the loss function parameters in this file.

In the directory ~/dpa_pretrain/src/utils/metric_learning_models.py

Model Evaluation

ROC-AUC,AUPR,F1max,Accuracy