Add MakHamTTS

README.md

@@ -38,6 +38,16 @@ from pythaitts import TTS
 tts = TTS(pretrained="fastthaig2p")
 file = tts.tts("สวัสดีครับ", speaker_idx="thai_som", filename="output.wav")
 
+# Use MakHamTTS Paradee (standalone, single-voice 8MB quantized INT8 ONNX model, offline without download)
+# MakHamTTS Paradee Sample Rate is 24000 Hz
+tts = TTS(pretrained="makham_paradee")
+file = tts.tts("สวัสดีค่ะ", filename="output.wav")
+
+# Use MakHamTTS (Kokoro-82M Thai base model, voices: newvoice_female, newvoice_male)
+# MakHamTTS Sample Rate is 24000 Hz
+tts = TTS(pretrained="makham")
+file = tts.tts("สวัสดีค่ะ", speaker_idx="newvoice_female", filename="output.wav")
+
 # Use Lunarlist ONNX
 # Sample Rate is 22050 Hz
 tts = TTS(pretrained="lunarlist_onnx")

demo_realtime.py

@@ -23,7 +23,7 @@ def main():
     # 1. Initialize TTS model
     print("Initializing FastThaiG2P TTS model...")
     t_start = time.time()
-    tts = TTS(pretrained="fastthaig2p")
+    tts = TTS(pretrained="makham_paradee")
     print(f"✓ Model loaded in {time.time() - t_start:.2f}s")
     print()
 

docs/index.rst

@@ -40,6 +40,16 @@ PyThaiTTS provides English-to-Thai-IPA Grapheme-to-Phoneme (G2P) conversion for
 
 .. autofunction:: en_ipa_to_thai_ipa
 
+MakHamTTS Models
+----------------
+MakHamTTS models powered by Kokoro-82M Thai and Paradee 1.1 distillation.
+
+.. autoclass:: MakHamTTS
+   :members:
+
+.. autoclass:: MakHamTTSParadee
+   :members:
+
 FastThaiG2P Components
 ----------------------
 

example_realtime.py

@@ -14,7 +14,7 @@ def main():
     print("-" * 50)
 
     # Initialize FastThaiG2P TTS
-    tts = TTS(pretrained="fastthaig2p")
+    tts = TTS(pretrained="makham_paradee")
 
     # 1. Stream speech from text string chunk-by-chunk
     text = "สวัสดีครับ ยินดีต้อนรับสู่ระบบสังเคราะห์เสียงภาษาไทยแบบเรียลไทม์"

pythaitts/__init__.py

@@ -13,6 +13,10 @@
     expand_maiyamok,
     load_custom_dict,
 )
+from pythaitts.pretrained.makham import (
+    MakHamTTS,
+    MakHamTTSParadee,
+)
 from pythaitts.pretrained.fastthaig2p.en_g2p import (
     EnglishG2P,
     en_g2p,
@@ -40,10 +44,10 @@ def __init__(
         **kwargs,
     ) -> None:
         """
-        :param str pretrained: TTS pretrained (fastthaig2p, lunarlist_onnx, khanomtan, lunarlist, vachana)
-        :param str mode: pretrained mode (fastthaig2p, lunarlist_onnx, and vachana don't support)
+        :param str pretrained: TTS pretrained (fastthaig2p, makham, makham_paradee, lunarlist_onnx, khanomtan, lunarlist, vachana)
+        :param str mode: pretrained mode (fastthaig2p, makham, makham_paradee, 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.)
+        :param str device: device for running model. (fastthaig2p, makham, makham_paradee, lunarlist_onnx, and vachana support CPU only.)
         :param custom_dict: Optional custom word dictionary for custom Thai pronunciation in words
 
         **Options for mode**
@@ -65,6 +69,9 @@ def __init__(
 
         For fastthaig2p tts model, \
         You can see more about fastthaig2p at `https://github.com/awslabs/FastThaiG2P <https://github.com/awslabs/FastThaiG2P>`_
+
+        For makham and makham_paradee tts models, \
+        You can see more about MakHamTTS at `https://huggingface.co/pythainlp/MakHamTTS <https://huggingface.co/pythainlp/MakHamTTS>`_
         """
         self.pretrained = pretrained
         self.mode = mode
@@ -93,6 +100,12 @@ def load_pretrained(self, version="1.0", **kwargs):
         elif self.pretrained in ("fastthaig2p", "FastThaiG2P"):
             from pythaitts.pretrained.fastthaig2p import FastThaiG2P
             self.model = FastThaiG2P(device=self.device, **kwargs)
+        elif self.pretrained in ("makham", "MakHamTTS"):
+            from pythaitts.pretrained.makham import MakHamTTS
+            self.model = MakHamTTS(**kwargs)
+        elif self.pretrained in ("makham_paradee", "MakHamTTS_paradee"):
+            from pythaitts.pretrained.makham import MakHamTTSParadee
+            self.model = MakHamTTSParadee(**kwargs)
         else:
             raise NotImplementedError(
                 "PyThaiTTS doesn't support %s pretrained." % self.pretrained
@@ -169,7 +182,7 @@ def tts(
         if self.pretrained == "lunarlist" or self.pretrained == "lunarlist_onnx":
             return self.model(text=text, return_type=return_type, filename=filename)
         elif self.pretrained == "vachana":
-            if speaker_idx in ("thai_som", "Linda", None):
+            if speaker_idx in ("thai_som", "Linda", "newvoice_female", None):
                 speaker_idx = "th_f_1"
             return self.model(
                 text=text,
@@ -179,7 +192,7 @@ def tts(
                 **kwargs,
             )
         elif self.pretrained in ("fastthaig2p", "FastThaiG2P"):
-            if speaker_idx in ("Linda", None):
+            if speaker_idx in ("Linda", "newvoice_female", None):
                 speaker_idx = "thai_som"
             call_kwargs = dict(kwargs)
             if effective_custom_dict is not None:
@@ -191,7 +204,31 @@ def tts(
                 filename=filename,
                 **call_kwargs,
             )
-        if speaker_idx in ("thai_som", None):
+        elif self.pretrained in ("makham", "MakHamTTS"):
+            if speaker_idx in ("Linda", "thai_som", None):
+                speaker_idx = "newvoice_female"
+            call_kwargs = dict(kwargs)
+            if effective_custom_dict is not None:
+                call_kwargs["custom_dict"] = effective_custom_dict
+            return self.model(
+                text=text,
+                speaker_idx=speaker_idx,
+                return_type=return_type,
+                filename=filename,
+                **call_kwargs,
+            )
+        elif self.pretrained in ("makham_paradee", "MakHamTTS_paradee"):
+            call_kwargs = dict(kwargs)
+            if effective_custom_dict is not None:
+                call_kwargs["custom_dict"] = effective_custom_dict
+            return self.model(
+                text=text,
+                speaker_idx=speaker_idx,
+                return_type=return_type,
+                filename=filename,
+                **call_kwargs,
+            )
+        if speaker_idx in ("thai_som", "newvoice_female", None):
             speaker_idx = "Linda"
         return self.model(
             text=text,
@@ -230,7 +267,7 @@ def stream(
         )
 
         if self.pretrained in ("fastthaig2p", "FastThaiG2P"):
-            if speaker_idx in ("Linda", None):
+            if speaker_idx in ("Linda", "newvoice_female", None):
                 speaker_idx = "thai_som"
             return self.model.stream(
                 text=text,
@@ -242,6 +279,30 @@ def stream(
                 custom_dict=effective_custom_dict,
                 **kwargs,
             )
+        elif self.pretrained in ("makham", "MakHamTTS"):
+            if speaker_idx in ("Linda", "thai_som", None):
+                speaker_idx = "newvoice_female"
+            return self.model.stream(
+                text=text,
+                speaker_idx=speaker_idx,
+                return_type=return_type,
+                play=play,
+                preprocess=preprocess,
+                max_phonemes=max_phonemes,
+                custom_dict=effective_custom_dict,
+                **kwargs,
+            )
+        elif self.pretrained in ("makham_paradee", "MakHamTTS_paradee"):
+            return self.model.stream(
+                text=text,
+                speaker_idx=speaker_idx,
+                return_type=return_type,
+                play=play,
+                preprocess=preprocess,
+                max_phonemes=max_phonemes,
+                custom_dict=effective_custom_dict,
+                **kwargs,
+            )
         else:
             from pythaitts.realtime import stream_text_to_chunks
 
@@ -276,6 +337,8 @@ def _gen():
     "en_g2p",
     "en_to_ipa",
     "en_ipa_to_thai_ipa",
+    "MakHamTTS",
+    "MakHamTTSParadee",
     "preprocess_text",
     "load_custom_dict",
     "num_to_thai",

pythaitts/pretrained/makham/__init__.py

@@ -0,0 +1,8 @@
+# -*- coding: utf-8 -*-
+"""
+MakHamTTS Models
+"""
+from pythaitts.pretrained.makham.makham_tts import MakHamTTS
+from pythaitts.pretrained.makham.paradee_tts import MakHamTTSParadee
+
+__all__ = ["MakHamTTS", "MakHamTTSParadee"]

pythaitts/pretrained/makham/data/config.json

@@ -0,0 +1,177 @@
+{
+  "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
+  },
+  "language": "th",
+  "model_type": "paradee-1.1",
+  "num_parameters": 7877918,
+  "architecture": "Paradee 1.1 (TextStudent s+mlp + VDecoder vhead v2)"
+}