Day2

Rohit
Dec 26, 2018

And here comes the second rung. Today, continuing the journey with motivated mind and full energy i wrote my piece of code.

Pledge

I, Rohit Sethi swear to complete #100DaysOfCode to the best of my ability and with true spirit.


Task Completed

On Day 2, I did the following :-

Trained a BiLSTM on Toxic comment classification 
and achieved a validation accuracy of 98.20%
using pre trained word embeddings 
which countered for the overfitting of baseline model.


Learnings

On Day 2, I learnt the following :-

  1. In detail study of word embeddings such as word2vec, Glove, fasttext and using their pre-trained embeddings directly.

  2. dropout and recurrent_dropout parameter in LSTM.

  3. Training BiLSTM with keras.

  4. Visualizing accuracy, loss, val_accuracy, val_loss with plotting after each epoch.


link to the work on Day 2 here