Corrections and improvements to README.md

[mouradvt04]

Few comments:

1. Add a short explanation on how to create matches_csv_file_name from the files (train, test, valid) that are in https://github.com/anhaidgroup/deepmatcher/blob/master/Datasets.md, concat all, keep line with label "1", and remove last column "label"

2. Correct vector_pairing_model --> topK_vector_pairing_model or vice-versa

3. Add line for cols_to_block = ["title","manufacturer","price"] with short explanation

4. Mention that configuration should be checked and that the embeddings file obtained from https://fasttext.cc/docs/en/pretrained-vectors.html is already in the correct folder

5. In the readme file, put "installation" before the 30 sec section