Converting sparsematrix to hdf5 is taking too much time even in Vaex and memory crashes

Question

I have a dataframe that contains text data and numerical features. I have vectorized text data and I plan to concatenate it with the remaining numerical data for running on