pass cache_dir to speech_separation pipeline

pyannote/audio/pipelines/speech_separation.py

@@ -27,6 +27,7 @@
 import math
 import textwrap
 import warnings
+from pathlib import Path
 from typing import Callable, Optional, Text, Tuple, Union
 
 import numpy as np
@@ -88,6 +89,9 @@ class SpeechSeparation(SpeakerDiarizationMixin, Pipeline):
         When loading private huggingface.co models, set `use_auth_token`
         to True or to a string containing your hugginface.co authentication
         token that can be obtained by running `huggingface-cli login`
+    cache_dir: Path or str, optional
+        Path to model cache directory. Defaults to content of PYANNOTE_CACHE
+        environment variable, or "~/.cache/torch/pyannote" when unset.
 
     Usage
     -----
@@ -132,11 +136,12 @@ def __init__(
         segmentation_batch_size: int = 1,
         der_variant: Optional[dict] = None,
         use_auth_token: Union[Text, None] = None,
+        cache_dir: Union[Path, Text, None] = None,
     ):
         super().__init__()
 
         self.segmentation_model = segmentation
-        model: Model = get_model(segmentation, use_auth_token=use_auth_token)
+        model: Model = get_model(segmentation, use_auth_token=use_auth_token, cache_dir=cache_dir)
 
         self.segmentation_step = segmentation_step
 
@@ -173,7 +178,7 @@ def __init__(
 
         else:
             self._embedding = PretrainedSpeakerEmbedding(
-                self.embedding, use_auth_token=use_auth_token
+                self.embedding, use_auth_token=use_auth_token, cache_dir=cache_dir
             )
             self._audio = Audio(sample_rate=self._embedding.sample_rate, mono="downmix")
             metric = self._embedding.metric