Fix typos in README

README.md

@@ -24,9 +24,9 @@ The model could only be trained for about `10%` of the whole dataset due to time
 
 ## Preprocessing and the tokenizer
 
-We tried to keep the preprocessing to the bare minimum. We ony replaced URLs, emails and social media user mentions with fixed tokens.
+We tried to keep the preprocessing to a bare minimum. We only replaced URLs, emails and social media user mentions with fixed tokens.
 
-Contrary to other pretrained Arabic LMs, we decided to not strip the Arabic diacritics and to keep them in the vocabulary.
+Contrary to other pretrained Arabic LMs, we decided to not strip the Arabic diacritics and to keep them part of the vocabulary.
 
 The tokenizer was trained on `5%` of the training set, with a vocabulary size of `64'000`.