README

1. Set up the env

We first need to create the env by

conda env create -f environment.yml

Then activate the env by

conda activate CMM

2. Train the model yourself

Please use train_model.py to train the model.

Note that I did not specify the device in that file.

To train the model python train_model --data_path DATA_path(i.e. BR0_data.csv) --meta_path(i.e. BR0_meta.csv)

3. Use pretrained model

One may use the pretrained scaler and model (trained over 10,000 epochs):

1. Get the scaler and model params from CMM DAE
2. Use the following instructions (please modify the path): python getData --data_path .\BR0_data.csv --scaler_path .\scaler.joblib --model_path .\DAE.pth --output_path D:\cm\CMM-Proj
3. We will get output.npy as the output, whose dims is (input.shape[0], 128).