Text summarization using Seq2Seq model in TensorFlow
Text Summarization or many other apps that get sequence input and return sequence outputs too are done using Seq2Seq models.
To know how it is done, please go and checkout my article on medium:
https://medium.com/@YasinShafiei/text-summarization-with-deep-learning-python-with-tensorflow-d0f3e329c3d2
The model made prediction on the test data and these are some of it's results:
Review: delightful item place bottom champagne added sweet twist holiday party made drink look wonderful loved well priced Original summary: price ever Predicted summary: delicious
Review: found health stores markets shop gotten reg pretzels brand sourdough taste sourdough flavor much salt excited find getting bit disappointment also label says spelt concern mine eating wheat spelt pretzels Original summary: great local Predicted summary: good pretzels
Review: assuming reading review realize instant coffee know means us like prefer ease instant coffee make strong weak like mint chocolate good job masking fact instant coffee one potentially negative comment would mint fresh strongest flavors Original summary: little nutritious little water way go better Predicted summary: it is ok
Review: almond accents oz bags used sold costco stopped carrying happy sold amazon meeting price quality requirements Original summary: olive texture Predicted summary: great product
Review: would preferred expiration date bit farther along five months considering buying six jars future going wait find stuff sale local store buy Original summary: cool occasions Predicted summary: good product