group_instructions.txt

######################################

For members of EECS 545 group wondering how to start, please look through the following instructions.

######################################

In general, 3 main files are used for data preparation, training and testing

File 1: data_preparation.ipynb
- Assumes that you have gone through the youtube video https://www.youtube.com/watch?v=M3ZWfamWrBM
- Basically, you will need to create folders to store DICOMM series data from NIFTI data
- After creating the DICOMM series data, you will use the tool to subsegment the data into equal size TestVolumes
- Subsegmented data is then coverted back into NIFTI files
- NIFTI data is then deployed into training and testing data sets to be used for training
- You can follow through the script as you watch through the video. The script provides some useful tools to help with your work
- Script also provides method to backtrack as required
- At the end, you should have a folder structure as defined in 'data/readme.txt'. You can also refer to that file to see
  the description of each folder in 'data'

File 2: train.py
- You can read through 'train.py' to understand how training is done
- Run file using command $ python train.py 
- Note: Takes really long using even on laptop with good GPU (7 hrs+ on laptop with RTX 3070 TI)

File 3: testing.ipynb
- File to run testing and inference of your trained model
- Does have repeated steps in File 2 especially in defining the preprocessing step and the model (Note: these must be the same
  as those in the training step for the results to be consistent)
- Go through file to see how to pull training and testing metrics, as well as to run inference