# Migration guide

## Basic workflow

<!-- import numpy as np
from sklearn.ensemble import RandomForestClassifier

X = np.array([[1, 2], [-1, 2]], dtype="float32")
y = np.array([0, 1], dtype="int32")

skl_model = RandomForestClassifier(n_estimators=1, max_depth=1)
skl_model.fit(X, y) -->

Call [`load_model()`](../python_api/index.html.md#nvforest.load_model), [`load_from_sklearn()`](../python_api/index.html.md#nvforest.load_from_sklearn),
or [`load_from_treelite_model()`](../python_api/index.html.md#nvforest.load_from_treelite_model). Note that it is no longer
necessary to specify the `is_classifier` parameter.

```python
# BEFORE
import cuml
fil_model = cuml.fil.ForestInference.load_from_sklearn(skl_model, is_classifier=True)
fil_model.optimize(batch_size=1024)
predictions = fil_model.predict(X)
probabilities = fil_model.predict_proba(X)
per_tree_pred = fil_model.predict_per_tree(X)
lead_ids = fil_model.apply(X)
```

```python
# AFTER
import nvforest
nvforest_model = nvforest.load_from_sklearn(skl_model)
nvforest_model = nvforest_model.optimize(batch_size=1024)
predictions = nvforest_model.predict(X)
probabilities = nvforest_model.predict_proba(X)
per_tree_pred = nvforest_model.predict_per_tree(X)
lead_ids = nvforest_model.apply(X)
```

## Device selection

Specify the `device` parameter when calling [`load_model()`](../python_api/index.html.md#nvforest.load_model).

```python
# BEFORE
with cuml.fil.set_fil_device_type("cpu"):
    fil_model = cuml.fil.ForestInference.load_from_sklearn(skl_model)
```

```python
# AFTER
nvforest_model = nvforest.load_from_sklearn(skl_model, device="cpu")
```

nvForest also differs from FIL when it comes to the behavior when no device is explicitly
specified. The `device` parameter defaults to `"auto"`. nvForest will attempt to
load the tree model onto a GPU device, if one is available. If no GPU is available,
nvForest will fall back to the CPU.