master: Text Summarization.

Author: purvasingh96

Chapter-wise code/Code - PyTorch/7. Attention Models/Google - Trax/1. Trax - Basics.ipynb renamed to Chapter-wise code/Code - PyTorch/7. Attention Models/1. NMT/Google - Trax/1. Trax - Basics.ipynb

@@ -0,0 +1,10 @@
+# Transformer Models
+
+## What was wrong with Seq2Seq Models?
+
+1. No parallel computaions. For longer sequence of text, a seq2seq model will take more number of timesteps to complete 
+the translation and as we know, with large sequences, the information tends to get lost in the network (vanishing gradient).
+LSTMs and GRUs can help to overcome the vanishing gradient problem, but even those will fail to process long sequences.<br><br>
+<img src="./images/1. drawbacks of seq2seq.png" width="50%"></img><br>
+
+2.