Fixed a typo

src/train_tokeniser.py

@@ -10,7 +10,7 @@ corpus = []
 
 with open("src/data/clean_corpus.jsonl", "r", encoding = "utf-8") as f_in:
     for i, line in enumerate(f_in):
-        if i >= 100000:  # Take only 100000 records for the tokeniser (no need to load everything in the corpus)
+        if i >= 100000:  # take only 100000 records for the tokeniser (no need to load everything in the corpus)
             break
 
         item = json.loads(line)
@@ -27,7 +27,7 @@ tokeniser = Tokenizer(models.BPE(unk_token = "<unk>"))
 # Normalisation (NFD and StripAccents are not used due to characters with diacritics, for instance)
 tokeniser.normalizer = Lowercase()
 
-# Basic whitespace pre-tokenization
+# Basic whitespace pre-tokenisation
 tokeniser.pre_tokenizer = Whitespace()
 
 # Trainer