Daml

🌟 **Feature Release**
- `2ed88a07` - Implement Drift Detection Metrics

This change adds 4 types of Drift Detection metrics which allow for the detection of potential drift in the dataset.

* Kolmogorov-Smirnov
* Cramér-von Mises
* Maximum Mean Discrepancy
* Classifier Uncertainty

The conceptual source is derived from [Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift](https://arxiv.org/abs/1810.11953) and the implementation is derived from [Alibi-Detect v0.11.4](https://github.com/SeldonIO/alibi-detect/tree/v0.11.4).

🚧 **Deprecations and Removals**
- `5cc48bec` - Divergence metric naming corrected to HP Divergence

Divergence metric output now returns a dictionary of `{ "divergence": float, "error": int }` instead of `{ "dpdivergence": float, "error": int }`. Code, documentation and tutorials have been updated to the correct nomenclature of HP (Henze-Penrose) divergence.

🌟 **Feature Release**
- `41b20d3a` - Add rules for release label pipeline workflow and merge request release template

🛠️ **Improvements and Enhancements**
- `7ee53c9c` - Update Divergence default to MST

🛠️ **Improvements and Enhancements**
- `1468aa5c` - Switch to markdown and updated docs

🛠️ **Improvements and Enhancements**
- `670a0db5` - Add support for classwise Sufficiency metrics
- `b96ee099` - Have sufficiency train and eval functions take indices and batch size instead of a DataLoader

🛠️ **Improvements and Enhancements**
- `5225c491` - Change output classes to dictionaries
- `45040682` - Make Sufficiency a stateful class and revise SufficiencyOutput
- `7c5fdcff` - Pass method as a parameter to determine metric algorithm to use
- `2e883f6d` - Add better optimizer to find global minimum
- `c3c78680` - Expose AETrainer to public API to use model multiple times after training

👾 **Fixes**
- `93564b95` - Updating pyproject.toml and lock file to set dependency less than numpy 2.0