maudiaz221/DQN-Trading-Bot

What it is

A Streamlit page that replays one stock through a bundled reinforcement-learning policy and charts the path. The visitor supplies Alpaca credentials, a ticker, a date range, and a bar size; the app loads historical bars from Alpaca's paper API into a discrete environment with sell, hold, and buy actions. A stable-baselines3 DQN checkpoint shipped with the space is applied next, and the page plots price, shares, and portfolio value with buy and sell marks. Researchers inspecting that checkpoint on their own history are the audience: the run only simulates fills and never sends broker orders, and the space card says DDPG while the code loads a DQN file.