Multinominal Logistic regression problem

Hello all,
I have a large datasets with 50000 rows with 10 different columns. This dataset has a column that has 100 different classes (categories as dependent variable) for classification. We are asked to find out if the given item belongs to which class (out of 100 categories).
How do you simplify the dataset? If I run a multinomial logistic regression, it takes a long time with many errors.
How do you perform any logistic regression or SVMs or Decision Tree in this case?

I would really appreciate your feedback!