Did my learner stop learning?


After lesson 6 (Part 1), I’ve been working on the NY City Taxi Fare dataset on Kaggle (only 1M rows out of 55M), and I’d stumbled upon few issues.


1> Why do I have the exact same MSE for all 5 epochs? Does that mean that the model had stopped learning after the first epoch? How do I get out of this situation?

2> I’m getting a MSE of 1.78 (RMSE - 1.33), and in the Kaggle leaderboard even the leader could only get 1.38 (albeit for the test data)!! Is it some kind of bug?! :innocent: