Save DDQN

[albert-z-guo]

Save DDQN

[Albert-Z-Guo]

Hi there, thanks for this feedback, I just updated the README and bumped TensorFlow to a supported stable version.

You can just run python3 evaluate.py --model_to_load=DDPG_ep10 to evaluate DDPG model. Model weights loading details are in DDPG.py.

Note that model weights may be broken when using a newer TensorFlow version so you may need to retrain the models.
In addition, TensorFlow has evolved a lot over the years and the way the models were written in this repo were largely deprecated and should be restructured accordingly.

This project was a toy project when I was learning deep reinforcement learning myself. For better framework and closer to real world trading applications, I'd recommend trying this repository: https://github.com/AI4Finance-Foundation/FinRL

Keep learning!