EnDeGen is a project that focuses on generating text using the Encoder-Decoder architecture with Recurrent Neural Networks (RNNs). It leverages the power of deep learning to predict and generate text sequences based on given input data.

EnDeGen is designed to generate text by learning patterns from a given dataset and predicting the next characters in a sequence. It utilizes the Encoder-Decoder model, which consists of two components:

• The CharacterLanguageModel class is responsible for encoding the input sequence. It takes the input characters, embeds them using an embedding layer, and passes them through a GRU (Gated Recurrent Unit) layer. The GRU layer processes the input sequence and returns the hidden states. Finally, the output of the GRU layer is passed through a dense layer to obtain the encoded representation of the input sequence.

• The OneStep class is responsible for decoding the encoded representation and generating the next character in the sequence. It takes the encoded representation, feeds it into the language model, and obtains the predicted logits for the next character. It then applies a temperature parameter to control the randomness of the predictions and applies a mask to prevent the generation of unknown characters. Finally, it samples from the predicted logits to generate the next character and returns it along with the model state.

[After training for 30 epochs, the model was able to generate the following text]

All: How nigh should do boys where no harm to keep Than younger conference was a point of your Boint powers. What, will our princely father York? Upon my love, think'st thou that I am little That go to chastity of our intone humour will church the mighty look od all in

Bolingbroke: Be it so I have the let morn of your having breath? Of high and lowly me! So, fellow, get me heaven, out o'er at Sicilia much in placent, all the prisoner to do fight: I do remember this be this: he's very little That wound this way how touch'd Slones, where you will have consumenes to yourself. You talk-worng, would wed both He thinks me down to such a party instinct, Travelling to practise himself be in love: I'll blush your honour.

ISABELLA: Which is too careless it and me? My grave be loss, that you hither hath a hand, And shows good words will maid to see thee dead, How doth the queen of Naples him. But let it serve to take our foes.

ANTONIO: I beseech your parish close by Caling home, That his bold hand.

For any questions or inquiries about the project, please reach out to LinkedIn

Happy text generation with EnDeGen!