/Transformer_classifier_pytorch

A notebook demonstrating use of Transformer's encoder block for classification using Pytorch only!

Primary LanguageJupyter Notebook

Transformer_classifier_pytorch

A notebook demonstrating use of Transformer's encoder block for multi-class classification using Pytorch only!

Open In Colab

  • It does not use any other library for the model (e.g. Hugging Face). Hence, only uses Pytorch.
  • Transformer itself is a sequence-to-sequence model consisting of Encoder and Decoder block.
  • However, in order to use Transformer for classification task we can only use Encoder block. This notebook demonstrate this!
  • Notebook is heavily commented with tensor's shape and possible explanation of logic.
  • Feel free to ask any question (raise issue) and contributions are also welcome regarding the notebook.

If you are interested in Neural Machine Translation with Transformer you can checkout my NMT repository.

Thanks