Coursera — Transfer Learning for NLP with TensorFlow Hub

Objectives → By end of this, you will be able to use

• Various Pre-trained NLP Models from TensorFlow Hub (thub.dev)
• Use transfer learning to fine tune text data
• Visualize Model Performance metric with Tensorboard

Image 01: Example of using pre-trained model in Keras

Problem Statement → Detect Toxic Comment (Quora Insincere Questions Classification)

!nvidia-smi → Command to tell GPU status

https://colab.research.google.com/drive/1o56qVx2GE7QR0g_uuC8LqPQqIW-cdK2q?authuser=3#scrollTo=PwWXcwno4wB9

Notebook on colab is opened using A/C → goelaakash12@gmail.com

import tensorflow as tf
print(“Version: “, tf.__version__)
print(“Hub version: “, hub.__version__)
print(“GPU is”, “available” if tf.config.list_physical_devices(‘GPU’) else “NOT AVAILABLE”)

Image 02: Sentence Representation

Image 03: Sentence Representation

TensorFlow Hub

SWIVEL — EMBEDDING (20 DIMENSION)

Paper → https://arxiv.org/abs/1602.02215

Quiz

END of QUIZ

References

https://www.coursera.org/learn/transfer-learning-nlp-tensorflow-hub/home/week/1