PPO model takes no positions during training (long-short setup)

Hi @LRAbbade 👋

I'm currently using the FinRL library in Google Colab for a class project, focusing on long-short trading using the PPO algorithm. I’ve built a custom environment similar to what FinRL uses, but the agent consistently avoids taking any positions during training, resulting in zero profit and zero rewards.

I noticed your recent commit related to YahooDownloader and wondered if you had any insight on how to best configure FinRL or PPO for long-short environments.

Would deeply appreciate any quick advice or pointer 🙏

Thanks for your great contributions to the repo!