- Database & SQL
- Tabular data & Traditional ML
- Datathon
- timeseries (stock prediction)
- Tensorflow Basics
- CV(Object Detection & Transferlearning)
- NLP Basics(Sentiment Analysis, Chatbot)
- NLP application(KoGPT : SFT, RM, PPO)
- Recommendation & Streamlit
- Mlops preview
- Final Project(with Donyang Systems)
- backbone : KoGPT-2
- Supervised Finetuning with kogpt-2 generated instruction dataset
- Make Reward Model with chatGPT, davinch, ada(good, bad, worst)
- Porximal Policy Optimization
- Reference : https://www.youtube.com/live/Iq8erq62s8c?si=wAIWcidtVbMi2-a2
- Data : MovieLens1M
- Model : AutoInt+
- Visualization : Streamlit
- Tensorflow, python, pandas