Dilyara Bareeva

> @annahedstroem Maybe it should be possible to pass parameter `gc_layer` to `quantus.evaluate` directly when using the GradCam method, like what is done in `quantus.explain` ? This would bypass the...

@vedal Thanks for your feedback! The purpose of `quantus.evaluate` is to compare multiple XAI methods by using multiple metrics, hence the `xai_methods` and `metrics` dictionaries. While it is possible to...

Lightning provide a "reproducibility" option for model training: https://lightning.ai/docs/pytorch/stable/common/trainer.html#reproducibility

TODO list to adjust benchmark: - [ ] Explicitly add HF dataset config name - [ ] Save and download model checkpoints separately from the main benchmark dictionary