Thanks for the comment, Nicholas! I understand your point but those graphs do have a purpose.

I used 20 percent of the S&P 500 data from the last ten years for validation. This means that the predictions for the last two years (2019 and 2020) on each of those graphs represent predictions on the validation data while the rest are predictions on the training data.

In other words, these graphs simultaneously display the model's performance on the training and validation data. The performance on the training data is close to perfect as expected, but this is not always the case with the validation data.