T

app.py

@@ -316,6 +316,59 @@ def build_ui():
                         prompt_audio_out = gr.Audio(label="生成音频", type="numpy")
                         prompt_status = gr.Textbox(label="状态", interactive=False)
                 prompt_btn.click(generate_voice_clone_from_prompt_file, inputs=[prompt_file, prompt_target_text, prompt_language], outputs=[prompt_audio_out, prompt_status])
+                gr.Markdown("---")
+
+                # Section 3: Traditional Voice Clone (Original)
+                gr.Markdown("### 3. 传统音色克隆（直接使用参考音频）")
+                gr.Markdown("直接上传参考音频生成语音（每次都需要提取特征）。")
+                with gr.Row():
+                    with gr.Column(scale=2):
+                        clone_ref_audio = gr.Audio(
+                            label="参考音频",
+                            type="numpy",
+                        )
+                        clone_ref_text = gr.Textbox(
+                            label="参考文本",
+                            lines=2,
+                            placeholder="输入参考音频中的确切文字...",
+                        )
+                        clone_xvector = gr.Checkbox(
+                            label="仅使用 x-vector",
+                            value=False,
+                        )
+
+                    with gr.Column(scale=2):
+                        clone_target_text = gr.Textbox(
+                            label="目标文本",
+                            lines=4,
+                            placeholder="输入要让克隆音色说话的文字...",
+                        )
+                        with gr.Row():
+                            clone_language = gr.Dropdown(
+                                label="语言",
+                                choices=LANGUAGES,
+                                value="Auto",
+                                interactive=True,
+                            )
+                            clone_model_size = gr.Dropdown(
+                                label="模型大小",
+                                choices=MODEL_SIZES,
+                                value="1.7B",
+                                interactive=True,
+                            )
+                        clone_btn = gr.Button("克隆并生成", variant="primary")
+
+                with gr.Row():
+                    clone_audio_out = gr.Audio(label="生成的音频", type="numpy")
+                    clone_status = gr.Textbox(label="状态", lines=2, interactive=False)
+
+                clone_btn.click(
+                    generate_voice_clone,
+                    inputs=[clone_ref_audio, clone_ref_text, clone_target_text, clone_language, clone_xvector, clone_model_size],
+                    outputs=[clone_audio_out, clone_status],
+                    api_name="generate_voice_clone"
+                )
+
     return demo
 
 if __name__ == "__main__":