Reproducing Figure 4 & Table 1 from FinRL: A Deep RL Library for Automated Stock Trading

Hi Dr. Liu and FinRL team,

Thanks so much for the great work on the FinRL library and your paper “FinRL: A Deep Reinforcement Learning Library for Automated Stock Trading in Quantitative Finance”. It’s been very helpful for my research on financial time series forecasting.

I’m currently trying to reproduce the single stock trading results using PPO, specifically the ones shown in Figure 4 and Table 1 of the paper. I noticed that while the GitHub repo is active, the Colab notebook for single stock trading seems to be broken or missing.

To better reproduce the results from the paper, I was hoping to clarify a few things:

1. What was the exact train/validation/test split used for the PPO single stock trading experiments in Figure 4? It seems different from the multi-stock setup shown in Figure 3.

2. Is the code for generating Figure 4 and Table 1 available anywhere, or would you be willing to share it by any chance?

3. If there’s an updated notebook or alternative setup for replicating the single stock PPO results, I’d be happy to try that out too.

Thanks again for all the great work, and I appreciate your help!