Perform classification on Image and float combined

Obviously, there is an ImageBlock to use for inputs that are images. I went through the other blocks in the docs and didn’t see a “FloatBlock” or “IntBlock”. How would I input a numeric value the same way I input an image in the datablock API? In my specific case, the value will come from a column in a dataframe.

I have an example here

We have a RegressionBlock() that you can use but it won’t show very well, my example shows a way to get .show_batch working well (to my knowledge, this could have been added)

1 Like

I should clarify a little more. Basically, I’m asking about a super simple multi-modal input scenario where the input is an image and a float value. The output is a classification (or regression possibly). I can write the network that can handle both inputs, I just need to create a dataloader (block?) that allows for both types of input.