Assignment

---

Take the code above, and make it work with any dataset. Submit the GitHub repo’s ReadMe file, where I can see answers to these questions:

1) what dataset you have used

2) what problem have you solved (fill in the blank, translation, text generation, etc)

3) the output of your training for 10 epochs

Solution

---

Link to GitHub Code

Link to Colab Code

1) Used Multik30 dataset

2) Problem Solved: Translation from French to English.

3) Training Logs

Epoch: 1, Train loss: 0.687, Val loss: 0.135, Epoch time = 79.095s
Epoch: 2, Train loss: 0.035, Val loss: 0.111, Epoch time = 78.633s
Epoch: 3, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.535s
Epoch: 4, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.522s
Epoch: 5, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.362s
Epoch: 6, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.350s
Epoch: 7, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.404s
Epoch: 8, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.544s
Epoch: 9, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.600s
Epoch: 10, Train loss: 0.000, Val loss: 0.111, Epoch time = 78.505s