Add MakHamTTS Paradee INT8 model as a built-in pretrained backend

README.md

@@ -38,6 +38,11 @@ from pythaitts import TTS
 tts = TTS(pretrained="fastthaig2p")
 file = tts.tts("สวัสดีครับ", speaker_idx="thai_som", filename="output.wav")
 
+# Use MakHamTTS (Paradee Thai INT8)
+# Sample Rate is 24000 Hz
+tts = TTS(pretrained="makhamtts")
+file = tts.tts("สวัสดีค่ะ ยินดีต้อนรับสู่บริการของเรา", speaker_idx="paradee_thai", filename="output.wav")
+
 # Use Lunarlist ONNX
 # Sample Rate is 22050 Hz
 tts = TTS(pretrained="lunarlist_onnx")

pythaitts/__init__.py

@@ -26,7 +26,7 @@
 class TTS:
     def __init__(self, pretrained="fastthaig2p", mode="last_checkpoint", version="1.0", device:str="cpu", **kwargs) -> None:
         """
-        :param str pretrained: TTS pretrained (fastthaig2p, lunarlist_onnx, khanomtan, lunarlist, vachana)
+        :param str pretrained: TTS pretrained (fastthaig2p, makhamtts, lunarlist_onnx, khanomtan, lunarlist, vachana)
         :param str mode: pretrained mode (fastthaig2p, lunarlist_onnx, and vachana don't support)
         :param str version: model version (default is 1.0 or 1.1)
         :param str device: device for running model. (fastthaig2p, lunarlist_onnx, and vachana support CPU only.)
@@ -72,6 +72,9 @@ def load_pretrained(self, version="1.0", **kwargs):
         elif self.pretrained == "vachana":
             from pythaitts.pretrained.vachana_tts import VachanaTTS
             self.model = VachanaTTS()
+        elif self.pretrained in ("makhamtts", "MakHamTTS"):
+            from pythaitts.pretrained.makham_tts import MakHamTTS
+            self.model = MakHamTTS()
         elif self.pretrained in ("fastthaig2p", "FastThaiG2P"):
             from pythaitts.pretrained.fastthaig2p import FastThaiG2P
             self.model = FastThaiG2P(device=self.device, **kwargs)
@@ -103,6 +106,10 @@ def tts(self, text: str, speaker_idx: str = "thai_som", language_idx: str = "th-
             if speaker_idx in ("thai_som", "Linda", None):
                 speaker_idx = "th_f_1"
             return self.model(text=text,speaker_idx=speaker_idx,return_type=return_type,filename=filename, **kwargs)
+        elif self.pretrained in ("makhamtts", "MakHamTTS"):
+            if speaker_idx in ("thai_som", "Linda", None):
+                speaker_idx = "paradee_thai"
+            return self.model(text=text,speaker_idx=speaker_idx,return_type=return_type,filename=filename, **kwargs)
         elif self.pretrained in ("fastthaig2p", "FastThaiG2P"):
             if speaker_idx in ("Linda", None):
                 speaker_idx = "thai_som"

pythaitts/pretrained/makham_tts.py

@@ -0,0 +1,78 @@
+# -*- coding: utf-8 -*-
+import json
+import tempfile
+from pathlib import Path
+
+import numpy as np
+import onnxruntime as ort
+import soundfile as sf
+
+from pythaitts.pretrained.fastthaig2p import G2P, ipa_to_kokoro
+
+
+_ASSET_DIR = Path(__file__).resolve().parent / "makhamtts"
+_DEFAULT_MODEL_PATH = _ASSET_DIR / "paradee_thai_int8.onnx"
+_DEFAULT_CONFIG_PATH = _ASSET_DIR / "config.json"
+
+
+class MakHamTTS:
+    SUPPORTED_VOICES = ["paradee_thai"]
+
+    def __init__(self, model_path=None, config_path=None) -> None:
+        self.model_path = Path(model_path) if model_path else _DEFAULT_MODEL_PATH
+        self.config_path = Path(config_path) if config_path else _DEFAULT_CONFIG_PATH
+
+        if not self.model_path.exists():
+            raise FileNotFoundError(f"MakHamTTS model file not found: {self.model_path}")
+        if not self.config_path.exists():
+            raise FileNotFoundError(f"MakHamTTS config file not found: {self.config_path}")
+
+        with open(self.config_path, "r", encoding="utf-8") as f:
+            cfg = json.load(f)
+
+        self.vocab = cfg["vocab"]
+        self.sample_rate = cfg.get("sample_rate", 24000)
+        self.g2p = G2P()
+        self.session = ort.InferenceSession(
+            str(self.model_path), providers=["CPUExecutionProvider"]
+        )
+
+    def tts(self, text: str, speed: float = 1.0):
+        phonemes = ipa_to_kokoro(self.g2p.convert(text))
+        ids = [0] + [self.vocab[c] for c in phonemes if c in self.vocab] + [0]
+        if len(ids) <= 2:
+            return np.array([], dtype=np.float32)
+        audio = self.session.run(
+            None,
+            {
+                "input_ids": np.array([ids], dtype=np.int64),
+                "speed": np.array([float(speed)], dtype=np.float32),
+            },
+        )[0][0]
+        return audio
+
+    def __call__(
+        self,
+        text: str,
+        speaker_idx: str = "paradee_thai",
+        return_type: str = "file",
+        filename: str = None,
+        **kwargs,
+    ):
+        if speaker_idx in ("thai_som", "Linda", None):
+            speaker_idx = "paradee_thai"
+        if speaker_idx not in self.SUPPORTED_VOICES:
+            raise ValueError(
+                f"Unsupported voice '{speaker_idx}'. Supported voices are: {', '.join(self.SUPPORTED_VOICES)}"
+            )
+
+        wav = self.tts(text, speed=kwargs.get("speed", 1.0))
+
+        if return_type == "waveform":
+            return wav
+        if filename is not None:
+            sf.write(filename, wav, self.sample_rate)
+            return filename
+        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
+            sf.write(fp.name, wav, self.sample_rate)
+            return fp.name

pythaitts/pretrained/makhamtts/config.json

@@ -0,0 +1,179 @@
+{
+  "istftnet": {
+    "upsample_kernel_sizes": [
+      20,
+      12
+    ],
+    "upsample_rates": [
+      10,
+      6
+    ],
+    "gen_istft_hop_size": 5,
+    "gen_istft_n_fft": 20,
+    "resblock_dilation_sizes": [
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ]
+    ],
+    "resblock_kernel_sizes": [
+      3,
+      7,
+      11
+    ],
+    "upsample_initial_channel": 512
+  },
+  "dim_in": 64,
+  "dropout": 0.2,
+  "hidden_dim": 512,
+  "max_conv_dim": 512,
+  "max_dur": 50,
+  "multispeaker": true,
+  "n_layer": 3,
+  "n_mels": 80,
+  "n_token": 178,
+  "style_dim": 128,
+  "text_encoder_kernel_size": 5,
+  "plbert": {
+    "hidden_size": 768,
+    "num_attention_heads": 12,
+    "intermediate_size": 2048,
+    "max_position_embeddings": 512,
+    "num_hidden_layers": 12,
+    "dropout": 0.1
+  },
+  "vocab": {
+    ";": 1,
+    ":": 2,
+    ",": 3,
+    ".": 4,
+    "!": 5,
+    "?": 6,
+    "—": 9,
+    "…": 10,
+    "\"": 11,
+    "(": 12,
+    ")": 13,
+    "“": 14,
+    "”": 15,
+    " ": 16,
+    "̃": 17,
+    "ʣ": 18,
+    "ʥ": 19,
+    "ʦ": 20,
+    "ʨ": 21,
+    "ᵝ": 22,
+    "ꭧ": 23,
+    "A": 24,
+    "I": 25,
+    "O": 31,
+    "Q": 33,
+    "S": 35,
+    "T": 36,
+    "W": 39,
+    "Y": 41,
+    "ᵊ": 42,
+    "a": 43,
+    "b": 44,
+    "c": 45,
+    "d": 46,
+    "e": 47,
+    "f": 48,
+    "h": 50,
+    "i": 51,
+    "j": 52,
+    "k": 53,
+    "l": 54,
+    "m": 55,
+    "n": 56,
+    "o": 57,
+    "p": 58,
+    "q": 59,
+    "r": 60,
+    "s": 61,
+    "t": 62,
+    "u": 63,
+    "v": 64,
+    "w": 65,
+    "x": 66,
+    "y": 67,
+    "z": 68,
+    "ɑ": 69,
+    "ɐ": 70,
+    "ɒ": 71,
+    "æ": 72,
+    "β": 75,
+    "ɔ": 76,
+    "ɕ": 77,
+    "ç": 78,
+    "ɖ": 80,
+    "ð": 81,
+    "ʤ": 82,
+    "ə": 83,
+    "ɚ": 85,
+    "ɛ": 86,
+    "ɜ": 87,
+    "ɟ": 90,
+    "ɡ": 92,
+    "ɥ": 99,
+    "ɨ": 101,
+    "ɪ": 102,
+    "ʝ": 103,
+    "ɯ": 110,
+    "ɰ": 111,
+    "ŋ": 112,
+    "ɳ": 113,
+    "ɲ": 114,
+    "ɴ": 115,
+    "ø": 116,
+    "ɸ": 118,
+    "θ": 119,
+    "œ": 120,
+    "ɹ": 123,
+    "ɾ": 125,
+    "ɻ": 126,
+    "ʁ": 128,
+    "ɽ": 129,
+    "ʂ": 130,
+    "ʃ": 131,
+    "ʈ": 132,
+    "ʧ": 133,
+    "ʊ": 135,
+    "ʋ": 136,
+    "ʌ": 138,
+    "ɣ": 139,
+    "ɤ": 140,
+    "χ": 142,
+    "ʎ": 143,
+    "ʒ": 147,
+    "ʔ": 148,
+    "ˈ": 156,
+    "ˌ": 157,
+    "ː": 158,
+    "ʰ": 162,
+    "ʲ": 164,
+    "↓": 169,
+    "→": 171,
+    "↗": 172,
+    "↘": 173,
+    "ᵻ": 177,
+    "↑": 170
+  },
+  "model_type": "paradee",
+  "name": "Paradee-Thai-Female",
+  "language": "th",
+  "sample_rate": 24000,
+  "teacher": "kokoro_thai",
+  "teacher_voice": "newvoice_female"
+}

setup.py

@@ -19,6 +19,7 @@
     packages=find_packages(),
     package_data={
         "pythaitts.pretrained.fastthaig2p": ["data/*", "data/fallback/*"],
+        "pythaitts.pretrained": ["makhamtts/*.onnx", "makhamtts/*.json"],
     },
     include_package_data=True,
     python_requires=">=3.6",

tests/test_makhamtts.py

@@ -0,0 +1,57 @@
+# -*- coding: utf-8 -*-
+import unittest
+from unittest.mock import Mock, patch
+
+import numpy as np
+
+from pythaitts import TTS
+
+
+class TestMakHamTTSIntegration(unittest.TestCase):
+    @patch("pythaitts.pretrained.makham_tts.MakHamTTS")
+    def test_makhamtts_model_initialization(self, mock_makham_cls):
+        mock_makham_cls.return_value = Mock()
+        tts = TTS(pretrained="makhamtts")
+        self.assertIsNotNone(tts.model)
+        self.assertEqual(tts.pretrained, "makhamtts")
+
+    @patch("pythaitts.pretrained.makham_tts.MakHamTTS")
+    def test_makhamtts_tts_call(self, mock_makham_cls):
+        mock_instance = Mock()
+        mock_instance.return_value = "/tmp/output.wav"
+        mock_makham_cls.return_value = mock_instance
+
+        tts = TTS(pretrained="makhamtts")
+        result = tts.tts("สวัสดีค่ะ", speaker_idx="paradee_thai", filename="/tmp/test.wav")
+
+        self.assertEqual(result, "/tmp/output.wav")
+        call_args = mock_instance.call_args
+        self.assertEqual(call_args.kwargs["text"], "สวัสดีค่ะ")
+        self.assertEqual(call_args.kwargs["speaker_idx"], "paradee_thai")
+        self.assertEqual(call_args.kwargs["filename"], "/tmp/test.wav")
+        self.assertEqual(call_args.kwargs["return_type"], "file")
+
+    @patch("pythaitts.pretrained.makham_tts.ipa_to_kokoro")
+    @patch("pythaitts.pretrained.makham_tts.G2P")
+    @patch("pythaitts.pretrained.makham_tts.ort.InferenceSession")
+    def test_makhamtts_waveform_return(
+        self, mock_session_cls, mock_g2p_cls, mock_ipa_to_kokoro
+    ):
+        from pythaitts.pretrained.makham_tts import MakHamTTS
+
+        mock_session = Mock()
+        mock_session.run.return_value = [np.array([[0.1, 0.2]], dtype=np.float32)]
+        mock_session_cls.return_value = mock_session
+        mock_g2p = Mock()
+        mock_g2p.convert.return_value = "/dummy/"
+        mock_g2p_cls.return_value = mock_g2p
+        mock_ipa_to_kokoro.return_value = "abc"
+
+        model = MakHamTTS()
+        wav = model("สวัสดีค่ะ", return_type="waveform")
+
+        self.assertTrue(np.array_equal(wav, np.array([0.1, 0.2], dtype=np.float32)))
+
+
+if __name__ == "__main__":
+    unittest.main()