kokoro 1.1-zh

kokoro/v11_zh/generate_v_11_zh.py

@@ -0,0 +1,101 @@
+from .generate_voices_bin import id2speaker
+
+
+def generate_kokoro_v_11():
+    s = f"""
+# kokoro-multi-lang-v1_1
+
+## Info about this model
+
+This model is kokoro v1.1-zh and it is from <https://huggingface.co/hexgrad/Kokoro-82M-v1.1-zh>
+
+It supports both `Chinese` and `English`.
+
+| Number of speakers | Sample rate |
+|--------------------|-------------|
+| {len(id2speaker)} | 24000|
+
+### Meaning of speaker prefix
+
+|Prefix|Meaning| sid range| Number of speakers|
+|---|---|---|---|
+|af | American female|0 - 1| 2|
+|bf| British female| 2| 1 |
+|zf| Chinese female| 3 - 57 | 55 |
+|zm| Chinese male| 58 - 102 | 45 |
+
+### speaker ID to speaker name (sid -> name)
+The mapping from speaker ID (sid) to speaker name is given below:
+
+"""
+
+    num_per_line = 4
+
+    t = "|" * (num_per_line + 2)
+    t += "\n"
+    t += "|---" * (num_per_line + 1)
+    t += "|\n"
+
+    t += f"|0 - {num_per_line-1}"
+
+    for i, (sid, name) in enumerate(id2speaker.items()):
+        t += f"|{sid} -> {name}"
+        if (i + 1) % num_per_line == 0:
+            t += "|\n"
+            if (i + 1) == min(i + num_per_line, len(id2speaker) - 1):
+                t += f"|{i+1}"
+            else:
+                t += f"|{i+1} - {min(i+num_per_line, len(id2speaker)-1)}"
+    while (i + 1) % num_per_line != 0:
+        t += "|"
+        i += 1
+    t += "|\n"
+
+    s += t
+
+    s += """
+### speaker name to speaker ID (name -> sid)
+The mapping from speaker name to speaker ID (sid) is given below:
+
+"""
+
+    t = "|" * (num_per_line + 2)
+    t += "\n"
+    t += "|---" * (num_per_line + 1)
+    t += "|\n"
+
+    t += f"|0 - {num_per_line-1}"
+
+    for i, (sid, name) in enumerate(id2speaker.items()):
+        t += f"|{name} -> {sid}"
+        if (i + 1) % num_per_line == 0:
+            t += "|\n"
+            t += f"|{i+1} - {min(i+num_per_line, len(id2speaker)-1)}"
+    while (i + 1) % num_per_line != 0:
+        t += "|"
+        i += 1
+    t += "|\n"
+
+    s += t
+
+    s += """
+## Samples
+
+For the following text:
+
+    This model supports both Chinese and English. 小米的核心价值观是什么？答案
+    是真诚热爱！有困难，请拨打110 或者18601200909。I am learning 机器学习.
+    我在研究 machine learning。What do you think 中英文说的如何呢?
+    今天是 2025年6月18号.
+
+sample audios for different speakers are listed below:
+
+"""
+    for sid, name in id2speaker.items():
+        s += f"\n### Speaker {sid} - {name}\n"
+        s += f"""\n<audio controls>
+  <source src="/sherpa/onnx/tts/all/kokoro/v1.1-zh/mp3//{sid}-{name}.mp3" type="audio/mp3">
+</audio>\n\n"""
+
+    with open("book/src/Chinese-English/kokoro-multi-lang-v1_1.md", "w") as f:
+        f.write(s)

kokoro/v11_zh/generate_voices_bin.py

@@ -0,0 +1,27 @@
+#!/usr/bin/env python3
+# Copyright    2025  Xiaomi Corp.        (authors: Fangjun Kuang)
+import torch
+from pathlib import Path
+
+
+speakers = [
+    "af_maple",
+    "af_sol",
+    "bf_vale",
+]
+
+d = Path(__file__).parent
+for i in range(1, 99 + 1):
+    name = "zf_{:03d}".format(i)
+    if Path(f"{d}/voices/{name}.txt").is_file():
+        speakers.append(name)
+
+for i in range(9, 100 + 1):
+    name = "zm_{:03d}".format(i)
+    if Path(f"{d}/voices/{name}.txt").is_file():
+        speakers.append(name)
+
+
+id2speaker = {index: value for index, value in enumerate(speakers)}
+
+speaker2id = {speaker: idx for idx, speaker in id2speaker.items()}

kokoro/v11_zh/update.sh

@@ -0,0 +1,5 @@
+#!/usr/bin/env bash
+set -ex
+
+cp /Users/fangjun/open-source/sherpa-onnx/scripts/kokoro/v1.1-zh/generate_voices_bin.py .
+