refactor: use external package for monotonic alignment

TTS/tts/utils/helpers.py

@@ -1,14 +1,10 @@
 import numpy as np
 import torch
+from monotonic_alignment_search import maximum_path as maximum_path_cython
 from scipy.stats import betabinom
 from torch.nn import functional as F
 
-try:
-    from TTS.tts.utils.monotonic_align.core import maximum_path_c
-
-    CYTHON = True
-except ModuleNotFoundError:
-    CYTHON = False
+CYTHON = True
 
 
 class StandardScaler:
@@ -174,23 +170,23 @@ def maximum_path(value, mask):
     return maximum_path_numpy(value, mask)
 
 
-def maximum_path_cython(value, mask):
-    """Cython optimised version.
-    Shapes:
-        - value: :math:`[B, T_en, T_de]`
-        - mask: :math:`[B, T_en, T_de]`
-    """
-    value = value * mask
-    device = value.device
-    dtype = value.dtype
-    value = value.data.cpu().numpy().astype(np.float32)
-    path = np.zeros_like(value).astype(np.int32)
-    mask = mask.data.cpu().numpy()
-
-    t_x_max = mask.sum(1)[:, 0].astype(np.int32)
-    t_y_max = mask.sum(2)[:, 0].astype(np.int32)
-    maximum_path_c(path, value, t_x_max, t_y_max)
-    return torch.from_numpy(path).to(device=device, dtype=dtype)
+# def maximum_path_cython(value, mask):
+#     """Cython optimised version.
+#     Shapes:
+#         - value: :math:`[B, T_en, T_de]`
+#         - mask: :math:`[B, T_en, T_de]`
+#     """
+#     value = value * mask
+#     device = value.device
+#     dtype = value.dtype
+#     value = value.data.cpu().numpy().astype(np.float32)
+#     path = np.zeros_like(value).astype(np.int32)
+#     mask = mask.data.cpu().numpy()
+
+#     t_x_max = mask.sum(1)[:, 0].astype(np.int32)
+#     t_y_max = mask.sum(2)[:, 0].astype(np.int32)
+#     maximum_path_c(path, value, t_x_max, t_y_max)
+#     return torch.from_numpy(path).to(device=device, dtype=dtype)
 
 
 def maximum_path_numpy(value, mask, max_neg_val=None):

pyproject.toml

@@ -64,6 +64,7 @@ dependencies = [
     # Coqui stack
     "coqui-tts-trainer>=0.1.4,<0.2.0",
     "coqpit>=0.0.16",
+    "monotonic-alignment-search>=0.1.0",
     # Gruut + supported languages
     "gruut[de,es,fr]>=2.4.0",
     # Tortoise