A curated collection of Deep Learning projects by Ashwin Kotha.
| Notebook | Description |
|---|---|
| Reuters_News_Topic_Classification_RNN.ipynb | Multi-class text classification on the Reuters dataset using SimpleRNN, LSTM, GRU, BiSimpleRNN, BiLSTM, and BiGRU architectures. Compares 6 RNN variants across 46 topic classes. |
| Reuters_Transformer_Text_Classification.ipynb | Transformer Encoder-based text classification on the Reuters newswire dataset. Implements a custom multi-head attention architecture with stacked encoder blocks for 46-class classification. |
- Python, TensorFlow / Keras
- NumPy, Matplotlib, Seaborn, Scikit-learn