This is the code for the "SiaDFP: A Disk Failure Prediction Framework Based on Siamese Neural Network in Large-Scale Data Center" and baselines.
The baselines of traditional machine learning models are based on the https://github.com/SidiLu001/disk_failure_prediction
Pre-request
CentOS 7.9.2009 python 3.7.6 PyTorch 1.2.0 CUDA Version: 10.2
Usage
- Clone the repo: https://github.com/fangInFBI/SiaDFP.git.
- Download the dataset from Dataset and put the data into the folder dataset/dataset1 and dataset/dataset2.
- Refer to the codes of corresponding sections for specific purposes.
Model
- loadData.py
The code to load the disk dataset.
- SiaDFPNet.py
The codes of the SiaDFP models in this paper.
- baselines.py
The codes of the baselines in this paper.
Train&&Test
- trainTest.py
The codes to train SiaDFP and evaluate the model on the test data.
- trainTestDLModels.py
The codes to train the baselines of traditional machine learning models and evaluate the models on the test data.
- trainTestMLModels.py
The codes to train the baselines of deep learning models and evaluate the models on the test data.