Most of the studies published on Text Summariza-tion are evaluated using ROUGE as a metric. Despite its popularity for Extractive Methods, its inability to capture semantic properties makes it less suitable for Abstractive Methods.
We all encountered a situation where Python was pushed to its limits.. This happens usually when we are dealing with heavy computation..
The answer is not "you open the refrigerator door, put the elephant inside, and close the door !" ; the answer is "Tiny Machine Learning !"