review of kokoro over

Author: Kamalabot

kokoro_review/app.py

@@ -0,0 +1,81 @@
+# /// script
+# requires-python = ">=3.12"
+# dependencies = [
+#     "gradio>=5.13.1",
+#     "kokoro-onnx>=0.3.8",
+#     "onnxruntime>=1.20.1",
+#     "colorlog>=6.9.0",
+#     "espeakng-loader>=0.2.4",
+#     "phonemizer-fork>=3.3.2",
+#     "numpy>=2.0.2",
+# ]
+#
+# ///
+
+"""
+wget https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/kokoro-v1.0.onnx
+wget https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/voices-v1.0.bin
+uv run examples/app.py
+"""
+
+import gradio as gr
+import numpy as np
+
+from kokoro_onnx import Kokoro
+from kokoro_onnx.tokenizer import Tokenizer
+
+tokenizer = Tokenizer()
+kokoro = Kokoro("kokoro-v1.0.onnx", "voices-v1.0.bin")
+
+
+SUPPORTED_LANGUAGES = ["en-us"]
+
+
+def create(text: str, voice: str, language: str, blend_voice_name: str = None):
+    phonemes = tokenizer.phonemize(text, lang=language)
+
+    # Blending
+    if blend_voice_name:
+        first_voice = kokoro.get_voice_style(voice)
+        second_voice = kokoro.get_voice_style(blend_voice_name)
+        voice = np.add(first_voice * (50 / 100), second_voice * (50 / 100))
+    samples, sample_rate = kokoro.create(
+        phonemes, voice=voice, speed=1.0, is_phonemes=True
+    )
+    return [(sample_rate, samples), phonemes]
+
+
+def create_app():
+    with gr.Blocks(theme=gr.themes.Soft(font=[gr.themes.GoogleFont("Roboto")])) as ui:
+        text_input = gr.TextArea(
+            label="Input Text",
+            rtl=False,
+            value="Kokoro TTS. Turning words into emotion, one voice at a time!",
+        )
+        language_input = gr.Dropdown(
+            label="Language",
+            value="en-us",
+            choices=SUPPORTED_LANGUAGES,
+        )
+        voice_input = gr.Dropdown(
+            label="Voice", value="af_sky", choices=sorted(kokoro.get_voices())
+        )
+        blend_voice_input = gr.Dropdown(
+            label="Blend Voice (Optional)",
+            value=None,
+            choices=sorted(kokoro.get_voices()) + [None],
+        )
+        submit_button = gr.Button("Create")
+        phonemes_output = gr.Textbox(label="Phonemes")
+        audio_output = gr.Audio()
+        submit_button.click(
+            fn=create,
+            inputs=[text_input, voice_input, language_input, blend_voice_input],
+            outputs=[audio_output, phonemes_output],
+        )
+    return ui
+
+
+ui = create_app()
+ui.launch(debug=True)
+