Скачать или смотреть RNNs Tutorial 3: Translating Sentences using an Encoder Decoder with Cross Attention

All of the code and diagrams for this tutorial can be found here: github.com/FilipStojanovicski/explaining_rnns_step_by_step

The content of this video was adapted from the PyTorch tutorial by Sean Robertson: https://pytorch.org/tutorials/interme...

Timeline
0:00 Intro
2:10 Architecture Diagram
11:39 Read Data
16:02 Encode Data into Numerical Format
30:02 Define Encoder
42:04 Define Decoder
56:22 Feed data through the Neural Network
1:05:25 Train the Neural Network
1:15:53 Make Predictions
1:33:29 Recap
1:36:24 Outro

In this video titled RNNs Tutorial 3: Translating Sentences using an Encoder Decoder with Cross Attention. We use everything we learned in the previous videos when we built RNNs for encoding the language of origin of surnames as well as an RNN for decoding in order to generate names in different languages. We combine the RNNs we built previously together and use them together in an Encoder Decoder architecture. We train this Neural Net in order to learn how to translate sentences from French to English. The encoder encodes the information about the french sentence which gets passed to the decoder which learns to generate the corresponding english sentence word by word. We walk through the entire process in code together. I hope you enjoy!

Join us for 'RNNs Tutorial 3: Translating Sentences using an Encoder Decoder with Cross Attention.' In this tutorial, we build on our previous RNN projects—encoding surnames by language of origin and generating names in different languages. Now, we combine these skills into an advanced Encoder-Decoder architecture to translate sentences from French to English. Watch as we train our neural network to encode French sentences and decode them into English, step-by-step. Perfect for those looking to enhance their NLP expertise. Dive into the code with us and master sentence translation with RNNs!