initial_tp passed to tiler.

minimal feature LSTM model built (without hyperparam finetuning and with minimal training - 10 epochs), saved as lstm_model.keras;

added descriptions for experimenting_w_MI.

benchmarked different SVM parameters and ways of gridsearch; benchmarked different imputation methods and their evaluation, thus finalising optimal way to run MI codes.

Got started with LSTM codes to replace SVM

uploaded data files for mutual information estimation; uploaded data for processing some data thru aliby and wela

committing jupyter notebooks for optimising SVM part of MI estimation, and using LSTM in place of SVM.

Also calculates cell_growth_rate.

Check h5 files before running postprocessor.