tensorlayer/seq2seq-chatbot

Implements a sequence-to-sequence encoder-decoder architecture with attention mechanisms for conversational AI, trained on Twitter and Cornell Movie Dialogue datasets. Built on TensorFlow 2.0+ with TensorLayer abstractions to reduce boilerplate while maintaining interpretability. Demonstrates end-to-end training pipeline including data preprocessing, model compilation, and multi-turn response generation from a compact LSTM-based implementation.

842 stars. No commits in the last 6 months.