Bugs found while working on slchallenge_demo notebook. (#392)

* WIP - Addressing bugs found while working on slchallenge_demo notebook.

* Adding `ids()` method to the dataset class fixed the duplicated ids problem.

* Adding some classifier evaluation metrics.

* Updated slchallenge notebook with confusion matrix plot.

* Trival commit

* Removing slchallenge_demo.ipynb from this PR. It is tracked in lincc-frameworks/notebooks_lf.

Author: drewoldag

src/hyrax/verbs/infer.py

@@ -112,7 +112,10 @@ def _save_batch(batch: Union[Tensor, list, tuple, dict], batch_results: Tensor):
                     object_ids[id] for id in range(write_index, write_index + len(batch_results))
                 ]
             elif batch_has_ids:
-                batch_object_ids = batch["object_id"].tolist()
+                if isinstance(batch["object_id"], list):
+                    batch_object_ids = batch["object_id"]
+                else:
+                    batch_object_ids = batch["object_id"].tolist()
             elif isinstance(batch, dict):
                 msg = "Dataset dictionary should be returning object_ids to avoid ordering errors. "
                 msg += "Modify the __getitem__ or __iter__ function of your dataset to include 'object_id' "