Added CLI beam width functionality

docs/api/linear.rst

@@ -32,6 +32,12 @@ The simplest usage is::
 
 .. autofunction:: get_positive_labels
 
+.. autoclass:: FlatModel
+   :members:
+
+.. autoclass:: TreeModel
+   :members:
+
 Load Dataset
 ^^^^^^^^^^^^
 

libmultilabel/linear/linear.py

@@ -17,10 +17,13 @@
     "predict_values",
     "get_topk_labels",
     "get_positive_labels",
+    "FlatModel",
 ]
 
 
 class FlatModel:
+    """A model returned from a training function."""
+
     def __init__(
         self,
         name: str,
@@ -619,7 +622,7 @@ def train_binary_and_multiclass(
 
 
 def predict_values(model, x: sparse.csr_matrix) -> np.ndarray:
-    """Calculates the decision values associated with x.
+    """Calculates the decision values associated with x, equivalent to model.predict_values(x).
 
     Args:
         model: A model returned from a training function.

libmultilabel/linear/tree.py

@@ -10,7 +10,7 @@
 
 from . import linear
 
-__all__ = ["train_tree"]
+__all__ = ["train_tree", "TreeModel"]
 
 
 class Node:
@@ -38,6 +38,8 @@ def dfs(self, visit: Callable[[Node], None]):
 
 
 class TreeModel:
+    """A model returned from train_tree."""
+
     def __init__(
         self,
         root: Node,

linear_trainer.py

@@ -19,9 +19,14 @@ def linear_test(config, model, datasets, label_mapping):
     else:
         labels = []
         scores = []
+
+    predict_kwargs = {}
+    if model.name == "tree":
+        predict_kwargs["beam_width"] = config.beam_width
+
     for i in tqdm(range(ceil(num_instance / config.eval_batch_size))):
         slice = np.s_[i * config.eval_batch_size : (i + 1) * config.eval_batch_size]
-        preds = linear.predict_values(model, datasets["test"]["x"][slice])
+        preds = model.predict_values(datasets["test"]["x"][slice], **predict_kwargs)
         target = datasets["test"]["y"][slice].toarray()
         metrics.update(preds, target)
         if k > 0: