# Common Modelling Use Cases

```{toctree}
:maxdepth: 1

notebooks/use_cases/preprocessing
notebooks/use_cases/autotune_scvi
notebooks/use_cases/minification
notebooks/use_cases/interpretability
notebooks/use_cases/multiGPU
notebooks/use_cases/VIVS_niche_gene_selection
```

```{customcard}
:path: notebooks/use_cases/preprocessing
:tags: Dev

Learn how to preprocess various types of data for use with scvi-tools models.
```

```{customcard}
:path: notebooks/use_cases/autotune_scvi
:tags: Analysis, Integration, Dev

Automatically find optimal set of hyperparameters using autotune.
```

```{customcard}
:path: notebooks/use_cases/minification
:tags: Minification, Dev

Minify a dataset by replacing count data with the model’s estimated parameters of the latent posterior distribution
```

```{customcard}
:path: notebooks/use_cases/interpretability
:tags: Analysis, Interpretability, Semi-supervised, Cell-type-annotation

Use integrated gradient or SHAP values for model explainability
```

```{customcard}
:path: notebooks/use_cases/multiGPU
:tags: Analysis

Example of how to train an SCVI model using multi GPU settings
```

```{customcard}
:path: notebooks/use_cases/VIVS_niche_gene_selection
:tags: Analysis, Gene-selection, Differential-comparison

Identify genes conditionally dependent on niche composition or another response using VIVS's conditional randomization test.
```