About this document
Optimal Hyperparameters For Deep LSTM-Networks For Sequence Labeling Tasks by Wardhana Halim Kusuma is a document available to read on EtoBox.
The document evaluates hyperparameters for deep LSTM networks on sequence labeling tasks. It tested over 50,000 network configurations on 5 tasks to determine which hyperparameters have the largest impact on performance, such as word embeddings and the last network layer. The optimized BiLSTM-CRF model is made publicly available.
- Author
- Wardhana Halim Kusuma
- Language
- EN