What are the details of the dataset you are trying to use for training (rows and columns)? This can sometimes occur if you have a very large dataset (particularly wide), even if you have de-selected features.
There is an improvement coming in the near future that improves the underlying dataset type and helps address this issue.
