Even after applying oversampling, classifier is still predicting only 1 class(majority class from before oversampling)

Question

I have a dataset that has 240 entries divided into 4 classes of 60 each. I am performing one vs all classification, so I take one class as class0 and all other 3 as class1.