Conversational-ChatBot

Phase 1 : Learning Phase(3 weeks)

• I learnt basics of machine learning, implemented gradient descent in python and understood the concepts involved.
• Learnt deep learning from online resources and textbooks.
• Understood how neural networks work, learnt feed forward neural nets and back propagation.
• Learnt concepts and uploaded the notes in git.
• Learnt aboubt natural language processing i.e., tokenization, embedding and about Reccurant neural nets, GRU and LSTM.
• Learnt how to implement these in tensorflow using keras api.

Phase 2 : Preprocessing(1 week)

• Collected the dataset, worked on preprocessing the dataset.
• Implented programs to convert speech to text and text to speech.

Phase 3 : Making and saving model(2 weeks)

• Learnt sequence to sequence modeling in natural language processing and implemented Encoder decoder model.
• Did hyper parameter tuning and tried to get best results for the given dataset.
• Saved the model in tflite model which can be interpreted in android studio.

Link to different models I tried : https://github.com/abhishek203/E-Abhishek

Phase 4 : App making(2 weeks)

• Learnt on android development.
• Made a chat app for our project.
• Did text preprocessing in java using the tokens imported as json files.
• Used tflite model to get the results.

Link to app files from android studio https://github.com/abhishek203/chatbot