← TensorFlow & Keras learning pathTensorFlow & Keras · Module 7 of 10
Sequences and text in Keras
Model text and time series.
📖 Key ideas
- TextVectorization turns raw text into token ids.
- Embedding layers learn word vectors.
- LSTM/GRU layers model sequences.
- Bidirectional wrappers read both directions.
- KerasNLP / KerasHub provide pretrained Transformers.
💡 Real-life example: TextVectorization is a dictionary turning each word into a page number the model can read.
TextVectorizationLayer converting text to ids
LSTMLong Short-Term Memory layer
BidirectionalProcesses sequence both ways
KerasHubPretrained model library