Update procedure/tips_and_tricks_for_training_with_qlora_on_cheap_desktop_PC.md

procedure/tips_and_tricks_for_training_with_qlora_on_cheap_desktop_PC.md

@@ -30,3 +30,4 @@ continous strings and if you train the model on raw book corpus, it will maybe a
 answer to question that could be answered by someone who read the book. For this, we need to transform the data from basic continous monologue to discussion where we have
 the name of the person who asks the question, followed by question, followed by newline and name of the person who responded to the question and an answer to the question.
 I've used this script as a base for my script https://pastebin.com/4DTKWpmY
+TODO