This is a beginner friendly repo for those who want to use DL techniques in Natural Language Processing for text classification.
It uses Tensorflow 2.0 framework for implementing deep learning models, with handy datasets.
You only need google account for the whole process to get started.Just open the notebooks using colab, which is an awesome tool provided from GOOGLE, cuz it gives you FREE GPUs!!!.
Only thing required is python, rest other things will be handled while exploration.
Knowing Tensorflow could help a lot to understand notebooks very quickly.
There are Various Classifier notebooks for different use-cases.
- IMDB-movie-review
- Sarcasm-Detection
- BBC (Text Classification into five topics{sport': 1, 'business': 2, 'politics': 3, 'tech': 4, 'entertainment': 5}, using BBC text-corpora).
- Handling text-data, which is far more difficult than handling images(Embeddings, Tokenization, Padding and Truncating, Stopwords and Lemmatization / Stemming)
- GRU, LSTM and other Recurrent Units, Bidirectional RNNs also 1d-convolutions, which is better technique for classification.
- Analyzing accuracy and loss.
These are very beginner friendly notebooks, so if you've any suggestion or edits, or you wanna add your ones. Just do PR :)
- Tensorflow.ai