Create a dataloader for text and other continuous and categorical column

Hey,

I have a dataset which contains both text and some tabular data. Can somebody guide me how can we create dataloader for this type of data?
@sgugger

Thanks.
Prakash Anand

You can create difference databunch for both text and tabular data. Then zip them up. This can make things a lot easier. To train the model you can copy some code from the fit_one_cycle and it would work fine.

1 Like

Can you provide some code with an example.

Just a normal databunch that you would create for text and tabular data. After that copy the code from the fit_one_cycle function. When the dataloader is called in the code just use both of your dataloaders.

1 Like