Description

A Python Keras API LSTM model integrated into a Discord API bot for text-to-emoji conversions. This project is my first attempt at a recurrent neural network (RNN).

• Respond 📢
• User says: The man goes fishing.
• Bot sends to the chat: 👨🎣
• Replace ✍️
• User says: I got a dog for christmas.
• Bot sends to the chat: I got a 🐶 for 🎄.
• React 🏷️
• User says: The police officer is really cool!
• Bot reacts to the message: 👮😎

Details 🤖

• TensorFlow LSTM Model
• Individual Project
• One Month Scope

Model's Features 📋

• Tokenized Descriptions
• Padded Sequences
• Scikit-Learn train_test_split
• TensorFlow Keras Model
• Adam Optimizer
• Trained On Over 700 Emojis

Model's Layers ⚙️

• Embedding for Input
• Long Short-Term Memory (LSTM) with 128 units
• Long Short-Term Memory (LSTM) with 56 units
• Dense for Output using Softmax Activiation

Bot's Features 🗂️

• Locally Hosted
• Command Event Handling
• On_Message Event Handling