Describe the bug

I was trying the new tiling pipeline of the version 1.2.0 and when trying to use a custom dataset and setting class path as anomalib.data.Folder and not as mvtec I was getting an error because of category, because there is no category in Folder class Category must be passed to the yaml file if not I am getting an error here in the creation of the ensemble environment category. My quick fix was to copy the code and change category with name (which is a must in Folder class and can function as the category in the ensemble engine).

Also in the docs, in the tile pipeline the accelerator was set as gpu while for me only cuda works: probably for this here

Dataset

Other (please specify in the text field below)

Model

PADiM

Steps to reproduce the behavior

just following the how to guide with a custom dataset

OS information

Python 3.10.12
anomalib 1.2.0

Expected behavior

I think that the pipeline should be able to get both folder and mvtec as an input in class path in the yaml configuration file and work accordingly

Screenshots

No response

Pip/GitHub

pip

What version/branch did you use?

No response

Configuration YAML

seed: 42
accelerator: "gpu"  # needed to be changed to cuda
default_root_dir: "results"
tiling:
  tile_size: [256, 256]
  stride: 256

normalization_stage: image 
thresholding:
  method: F1AdaptiveThreshold 
  stage: image 

data:
  class_path: anomalib.data.MVTec # error when I changed to anomalib.data.Folder
  init_args: ...

TrainModels:
    class_path: Padim

Logs

-

Code of Conduct

• I agree to follow this project's Code of Conduct