How to use pretrained embeddings in machine learning project?


I have an NLP project where I would like to compare the results of the NN (torchtext) approach to the classic ML approach in a way to use pre-trained glove/fasttext embeddings for ML, for example, Linear Regression in sklearn.
Is it possible? This approach is because I want to use the same dataset split and the same vectors.

Thank you for any help, especially code snippets or links with examples