feat: initial public release v0.1.0 — MLX port of pyannote-speaker-diarization-3.1

Byte-parity with pyannote-PyTorch reference (cosine 0.763718 identical
at 6 decimals on 200 cross-window slot pairs). 2.5x faster than
pyannote-MPS on Apple Silicon native.

Extracted from gitea.tavportal.com/olivier/MLX_CONVERTOR commit 5f9eafa.

tests/unit/test_diar_segmentation_shape.py

@@ -0,0 +1,9 @@
+import mlx.core as mx
+from pyannote_diarization_3_1_mlx.segmentation import SegmentationModel
+
+
+def test_segmentation_full_shape():
+    m = SegmentationModel()
+    x = mx.zeros((1, 1, 160000))  # 10s @ 16k mono
+    out = m(x)
+    assert out.shape == (1, 589, 7), f"got {out.shape}"