Understanding FinRL for special use case

[RaiAmanRai]

Hi, I am new to FinRL and RL itself. I have a special case of deep RL and I am not suer if it can be solved using FinRL.

So the problem is such that I have a time dependent multi-dimensional data which misses the ground truth labels, but I have a reward funcition instead.

Can I use FinRL to consume the data being generated in realtime and fit it using the reward function (reward function will take the predictions of previous time step, adjust weights of model and make new predictions)