This project aims to use deep reinforcement learning for asset allocation of 20 US stocks. We experiment with the A2C, PPO, DDPG agents from Stable Baseline3. We also build a policy gradient agent which achieves much higher returns than the uniform weight baseline. Some of the codes are inspired by the following github repository: /~https://github.com/Musonda2day/Asset-Portfolio-Management-usingDeep-Reinforcement-Learning- /~https://github.com/deepcrypto/Reinforcement-learning-in-portfolio-management-