Sagemaker canvas numeric prediction RSME values are high with same datasets used

I have created a canvas that uses the datasets below:

and I let it do the training using RSME, i did not change any advance settings, and the target column is price. What I am trying to achieve is that it can predict the price of the product. However, upon successfully build, the RSME is 0.769 (Standard Build), which is quite bad.

And when I request it to predict with the same exact datasets but without the price column, the predicted values for each of the products are way off from the actual. I am wondering is there any settings that I have missed or anything that I could do to have a better prediction accuracy and precision.

Example ; Left side is original datasets, right ride is predicted price from Canvas