My ML project for 6th Sem 😉 👌

This is a project which I took during 6th Semester to make a Recursive Neural Network(RNN) train on a textual dataset, and then use it to output real words, given some input. It uses a type of RNNs called Long Short Term Memory(LSTM) Networks to do the job.

Here are some useful pointers:

• Understanding LSTM Networks
• The Unreasonable Effectiveness of Recurrent Neural Networks

First of all, install these dependencies:

pip3 install numpy tensorflow

Then once loaded, run the bot using Python3

$ python3 bot.py

The bot connects to a channel on Freenode using a inbuilt Bot class. To send messages to the bot type this.

[insert-bot-name] hey

Auto generated messages using the bot's trained models are sent. Models are trained on the tensorflow Shakespeare text.

To train a new model, try changing this 2 lines:

EPOCHS=15

history = model.fit(dataset.repeat(), epochs=EPOCHS, steps_per_epoch=steps_per_epoch, callbacks=[checkpoint_callback])

EPOCHS being the number of times you want to train the model.