Add FastFlow model training testing inference via Anomalib API (#386)

* Refactored MVTec datamodule

* 🏷️  Rename BTech datamodule

* 🏷️  Rename Folder datamodule

* Create datamodule jupyter notebook

* Added mvtec into jupyter notebook

* Apply black formatter

* Finished MVTec

* 🏷 Rename the name of the folder

* 🏷  Renamed anomaly-datamodule.ipynb to mvtec.ipynb

* ➕ Created BTech notebook

* 🚚 Move the main description from mvtec to README.md

* Polish btech jupyter notebook

* ➕Created folder jupyter notebook

* Addres PR comments.

* Format the notebooks

* Added black, isort, flake8 and pylint

* ➕ Add mdformat to dev requirements

* 🛠  Update the nnotebook markdowns

* Configured pre-commit for notebooks

* ➕ Add DATAMODULE_REGISTRY to the datamodules.

* ➕ Added relative imports to __init__ modules.

* Register callbacks via @CALLBACK_REGISTRY

* ➕ Added missing callback imports

* ➕ Add AnomalibCLI

* ➕ Add AnomalibCLI

* ➕ Add `normalization_method` to `MetricsConfigurationCallback`

* ➕ Add padim config file

* 🛠 Fix pytorch-lightning version to add extra

* Update torchmetrics version requirement

* Add anomalib CLI entrypoint

* Remove --save_images flag from the cli

* Refactor trainer.py and create main() function

* Updated anomalib entrypoint. Removed trainer.py from cli

* Add logger to cli

* ➕ Add patchcore config

* ➕ Add Cflow CLI config

* 🚚  Move padim and patchcore configs to config directory

* Add cflow, dfkde, dfm, draem, fastflow, padim and patchcore

* ➕ Add ganomaly config

* add reverse distillation and stfpm configs

* Update notebooks

* Create the project directory only during traaining.

* Added notebooks

* 🏷 Renamed config directory to configs

* Fix the project directory creation logic.

* Add the training CLI command to the README.md

* 🛠  Fix incorrect statement.

* Fix patchcore training

* Added fastflow training

* Added fastflow jupyter notbeook

* Removeed models jupyter notebooks

* Removed trainer.py

* Addressed PR comments'

* Address PR comments.

* Address markdown checks

configs/model/fastflow.yaml

@@ -29,7 +29,7 @@ model:
     hidden_ratio: 1.0 # options: [1.0, 1.0, 0.16, 0.16] - for each supported backbone
 
 optimizer:
-  class_path: torch.optim.Adam
+  class_path: torch.optim._multi_tensor.Adam
   init_args:
     lr: 0.001
     weight_decay: 0.00001