In this project, I have used pretrainned models from Tensorflow hub with tf.kear for text classification and visualise the performace using TensorBoard.
"Quora Insincere Questions Classification" from kaggle will be used and our models will be trained to detect if the questions asked on Quora is sincere or not.
Five pre-trained models were used as stated below:
- gnews-swivel-20dim
- nnlm-en-dim50
- nnlm-en-dim128
- universal-sentence-encoder
- universal-sentence-encoder-large
To run the notebook, please make a copy on Google Colab first