Sequence-to-sequence models

Compute interpretation

Encoder-decoder sequence modeling that used GPU-parallel recurrent workloads before the Transformer made dense attention dominant.

Supporting reading cards

• Neural Machine Translation by Jointly Learning to Align and Translate (2014, single_gpu_deep_learning)
• Sequence to Sequence Learning with Neural Networks (2014, single_gpu_deep_learning)
• Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation (2016, multi_gpu_dense_training)

Obsolete or less central under later compute

Track this only through linked reading cards; do not treat this method page as standalone evidence.