Update main/app/app.py

main/app/app.py

@@ -74,6 +74,7 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
                     with gr.Accordion(translations["use_url"], open=False):
                         url = gr.Textbox(label=translations["url_audio"], value="", placeholder="https://www.youtube.com/...", scale=6)
                         download_button = gr.Button(translations["downloads"])
+            with gr.Row():
                 with gr.Column():
                     with gr.Row():
                         clean_strength = gr.Slider(label=translations["clean_strength"], info=translations["clean_strength_info"], minimum=0, maximum=1, value=0.5, step=0.1, interactive=True, visible=cleaner.value)
@@ -147,6 +148,7 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
                             model_index = gr.Dropdown(label=translations["index_path"], choices=index_path, value=index_path[0] if len(index_path) >= 1 else "", interactive=True, allow_custom_value=True)
                         refesh = gr.Button(translations["refesh"])
                     
+                    
             with gr.Row(): 
                 with gr.Column():
                     audio_select = gr.Dropdown(label=translations["select_separate"], choices=[], value="", interactive=True, allow_custom_value=True, visible=False)
@@ -432,149 +434,112 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
                     api_name="convert_audio"
                 )
 
-        with gr.TabItem(translations["convert_with_whisper"], visible=configs.get("convert_with_whisper", True)):
-            gr.Markdown(f"## {translations['convert_with_whisper']}")
+        
+        with gr.TabItem(translations["downloads"], visible=configs.get("downloads_tab", True)):
+            gr.Markdown(translations["download_markdown"])
             with gr.Row():
-                gr.Markdown(translations["convert_with_whisper_info"])
+                gr.Markdown(translations["download_markdown_2"])
             with gr.Row():
-                with gr.Column():
-                    with gr.Group():
-                        with gr.Row():
-                            cleaner2 = gr.Checkbox(label=translations["clear_audio"], value=False, interactive=True)
-                            autotune2 = gr.Checkbox(label=translations["autotune"], value=False, interactive=True)
-                            checkpointing2 = gr.Checkbox(label=translations["memory_efficient_training"], value=False, interactive=True)
-                            formant_shifting2 = gr.Checkbox(label=translations["formantshift"], value=False, interactive=True)
+                with gr.Accordion(translations["model_download"], open=True):
+                    with gr.Row():
+                        downloadmodel = gr.Radio(label=translations["model_download_select"], choices=[translations["download_url"], translations["download_from_csv"], translations["search_models"], translations["upload"]], interactive=True, value=translations["download_url"])
+                    with gr.Row():
+                        gr.Markdown("___")
+                    with gr.Column():
                         with gr.Row():
-                            num_spk = gr.Slider(minimum=2, maximum=8, step=1, info=translations["num_spk_info"], label=translations["num_spk"], value=2, interactive=True)
-            with gr.Row():
-                with gr.Column():
-                    convert_button3 = gr.Button(translations["convert_audio"], variant="primary")
+                            url_input = gr.Textbox(label=translations["model_url"], value="", placeholder="https://...", scale=6)
+                            download_model_name = gr.Textbox(label=translations["modelname"], value="", placeholder=translations["modelname"], scale=2)
+                        url_download = gr.Button(value=translations["downloads"], scale=2)
+                    with gr.Column():
+                        model_browser = gr.Dropdown(choices=models.keys(), label=translations["model_warehouse"], scale=8, allow_custom_value=True, visible=False)
+                        download_from_browser = gr.Button(value=translations["get_model"], scale=2, variant="primary", visible=False)
+                    with gr.Column():
+                        search_name = gr.Textbox(label=translations["name_to_search"], placeholder=translations["modelname"], interactive=True, scale=8, visible=False)
+                        search = gr.Button(translations["search_2"], scale=2, visible=False)
+                        search_dropdown = gr.Dropdown(label=translations["select_download_model"], value="", choices=[], allow_custom_value=True, interactive=False, visible=False)
+                        download = gr.Button(translations["downloads"], variant="primary", visible=False)
+                    with gr.Column():
+                        model_upload = gr.File(label=translations["drop_model"], file_types=[".pth", ".onnx", ".index", ".zip"], visible=False)
             with gr.Row():
-                with gr.Column():
-                    with gr.Accordion(translations["model_accordion"] + " 1", open=True):
-                        with gr.Row():
-                            model_pth2 = gr.Dropdown(label=translations["model_name"], choices=model_name, value=model_name[0] if len(model_name) >= 1 else "", interactive=True, allow_custom_value=True)
-                            model_index2 = gr.Dropdown(label=translations["index_path"], choices=index_path, value=index_path[0] if len(index_path) >= 1 else "", interactive=True, allow_custom_value=True)
-                        with gr.Row():
-                            refesh2 = gr.Button(translations["refesh"])
-                        with gr.Row():
-                            pitch3 = gr.Slider(minimum=-20, maximum=20, step=1, info=translations["pitch_info"], label=translations["pitch"], value=0, interactive=True)
-                            index_strength2 = gr.Slider(label=translations["index_strength"], info=translations["index_strength_info"], minimum=0, maximum=1, value=0.5, step=0.01, interactive=True, visible=model_index2.value != "")
-                    with gr.Accordion(translations["input_output"], open=False):
-                        with gr.Column():
-                            export_format2 = gr.Radio(label=translations["export_format"], info=translations["export_info"], choices=["wav", "mp3", "flac", "ogg", "opus", "m4a", "mp4", "aac", "alac", "wma", "aiff", "webm", "ac3"], value="wav", interactive=True)
-                            input_audio1 = gr.Dropdown(label=translations["audio_path"], value="", choices=paths_for_files, info=translations["provide_audio"], allow_custom_value=True, interactive=True)
-                            output_audio2 = gr.Textbox(label=translations["output_path"], value="audios/output.wav", placeholder="audios/output.wav", info=translations["output_path_info"], interactive=True)
-                        with gr.Column():
-                            refesh4 = gr.Button(translations["refesh"])
-                        with gr.Row():
-                            input2 = gr.File(label=translations["drop_audio"], file_types=[".wav", ".mp3", ".flac", ".ogg", ".opus", ".m4a", ".mp4", ".aac", ".alac", ".wma", ".aiff", ".webm", ".ac3"])
-                with gr.Column():
-                    with gr.Accordion(translations["model_accordion"] + " 2", open=True):
-                        with gr.Row():
-                            model_pth3 = gr.Dropdown(label=translations["model_name"], choices=model_name, value=model_name[0] if len(model_name) >= 1 else "", interactive=True, allow_custom_value=True)
-                            model_index3 = gr.Dropdown(label=translations["index_path"], choices=index_path, value=index_path[0] if len(index_path) >= 1 else "", interactive=True, allow_custom_value=True)
-                        with gr.Row():
-                            refesh3 = gr.Button(translations["refesh"])
-                        with gr.Row():
-                            pitch4 = gr.Slider(minimum=-20, maximum=20, step=1, info=translations["pitch_info"], label=translations["pitch"], value=0, interactive=True)
-                            index_strength3 = gr.Slider(label=translations["index_strength"], info=translations["index_strength_info"], minimum=0, maximum=1, value=0.5, step=0.01, interactive=True, visible=model_index3.value != "")
-                    with gr.Accordion(translations["setting"], open=False):
-                        with gr.Row():
-                            model_size = gr.Radio(label=translations["model_size"], info=translations["model_size_info"], choices=["tiny", "tiny.en", "base", "base.en", "small", "small.en", "medium", "medium.en", "large-v1", "large-v2", "large-v3", "large-v3-turbo"], value="medium", interactive=True)
-                        with gr.Accordion(translations["f0_method"], open=False):
-                            with gr.Group():
-                                with gr.Row():
-                                    onnx_f0_mode4 = gr.Checkbox(label=translations["f0_onnx_mode"], info=translations["f0_onnx_mode_info"], value=False, interactive=True)
-                                    unlock_full_method2 = gr.Checkbox(label=translations["f0_unlock"], info=translations["f0_unlock_info"], value=False, interactive=True)
-                                method3 = gr.Radio(label=translations["f0_method"], info=translations["f0_method_info"], choices=method_f0+["hybrid"], value="rmvpe", interactive=True)
-                                hybrid_method3 = gr.Dropdown(label=translations["f0_method_hybrid"], info=translations["f0_method_hybrid_info"], choices=["hybrid[pm+dio]", "hybrid[pm+crepe-tiny]", "hybrid[pm+crepe]", "hybrid[pm+fcpe]", "hybrid[pm+rmvpe]", "hybrid[pm+harvest]", "hybrid[pm+yin]", "hybrid[dio+crepe-tiny]", "hybrid[dio+crepe]", "hybrid[dio+fcpe]", "hybrid[dio+rmvpe]", "hybrid[dio+harvest]", "hybrid[dio+yin]", "hybrid[crepe-tiny+crepe]", "hybrid[crepe-tiny+fcpe]", "hybrid[crepe-tiny+rmvpe]", "hybrid[crepe-tiny+harvest]", "hybrid[crepe+fcpe]", "hybrid[crepe+rmvpe]", "hybrid[crepe+harvest]", "hybrid[crepe+yin]", "hybrid[fcpe+rmvpe]", "hybrid[fcpe+harvest]", "hybrid[fcpe+yin]", "hybrid[rmvpe+harvest]", "hybrid[rmvpe+yin]", "hybrid[harvest+yin]"], value="hybrid[pm+dio]", interactive=True, allow_custom_value=True, visible=method3.value == "hybrid")
-                            hop_length3 = gr.Slider(label="Hop length", info=translations["hop_length_info"], minimum=1, maximum=512, value=128, step=1, interactive=True, visible=False)
-                        with gr.Accordion(translations["hubert_model"], open=False):
-                            embed_mode3 = gr.Radio(label=translations["embed_mode"], info=translations["embed_mode_info"], value="fairseq", choices=embedders_mode, interactive=True, visible=True)
-                            embedders3 = gr.Radio(label=translations["hubert_model"], info=translations["hubert_info"], choices=embedders_model, value="hubert_base", interactive=True)
-                            custom_embedders3 = gr.Textbox(label=translations["modelname"], info=translations["modelname_info"], value="", placeholder="hubert_base", interactive=True, visible=embedders3.value == "custom")
-                        with gr.Column():      
-                            clean_strength3 = gr.Slider(label=translations["clean_strength"], info=translations["clean_strength_info"], minimum=0, maximum=1, value=0.5, step=0.1, interactive=True, visible=cleaner2.value)
-                            f0_autotune_strength3 = gr.Slider(minimum=0, maximum=1, label=translations["autotune_rate"], info=translations["autotune_rate_info"], value=1, step=0.1, interactive=True, visible=autotune.value)
-                            resample_sr3 = gr.Slider(minimum=0, maximum=96000, label=translations["resample"], info=translations["resample_info"], value=0, step=1, interactive=True)
-                            filter_radius3 = gr.Slider(minimum=0, maximum=7, label=translations["filter_radius"], info=translations["filter_radius_info"], value=3, step=1, interactive=True)
-                            volume_envelope3 = gr.Slider(minimum=0, maximum=1, label=translations["volume_envelope"], info=translations["volume_envelope_info"], value=1, step=0.1, interactive=True)
-                            protect3 = gr.Slider(minimum=0, maximum=1, label=translations["protect"], info=translations["protect_info"], value=0.5, step=0.01, interactive=True)
+                with gr.Accordion(translations["download_pretrained_2"], open=False):
+                    with gr.Row():
+                        pretrain_download_choices = gr.Radio(label=translations["model_download_select"], choices=[translations["download_url"], translations["list_model"], translations["upload"]], value=translations["download_url"], interactive=True)  
+                    with gr.Row():
+                        gr.Markdown("___")
+                    with gr.Column():
                         with gr.Row():
-                            formant_qfrency3 = gr.Slider(value=1.0, label=translations["formant_qfrency"] + " 1", info=translations["formant_qfrency"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
-                            formant_timbre3 = gr.Slider(value=1.0, label=translations["formant_timbre"] + " 1", info=translations["formant_timbre"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
+                            pretrainD = gr.Textbox(label=translations["pretrained_url"].format(dg="D"), value="", info=translations["only_huggingface"], placeholder="https://...", interactive=True, scale=4)
+                            pretrainG = gr.Textbox(label=translations["pretrained_url"].format(dg="G"), value="", info=translations["only_huggingface"], placeholder="https://...", interactive=True, scale=4)
+                        download_pretrain_button = gr.Button(translations["downloads"], scale=2)
+                    with gr.Column():
                         with gr.Row():
-                            formant_qfrency4 = gr.Slider(value=1.0, label=translations["formant_qfrency"] + " 2", info=translations["formant_qfrency"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
-                            formant_timbre4 = gr.Slider(value=1.0, label=translations["formant_timbre"] + " 2", info=translations["formant_timbre"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
-            with gr.Row():
-                gr.Markdown(translations["input_output"])
-            with gr.Row():
-                play_audio2 = gr.Audio(show_download_button=True, interactive=False, label=translations["input_audio"])
-                play_audio3 = gr.Audio(show_download_button=True, interactive=False, label=translations["output_file_tts_convert"])
-            with gr.Row():
-                autotune2.change(fn=visible, inputs=[autotune2], outputs=[f0_autotune_strength3])
-                cleaner2.change(fn=visible, inputs=[cleaner2], outputs=[clean_strength3])
-                method3.change(fn=lambda method, hybrid: [visible(method == "hybrid"), hoplength_show(method, hybrid)], inputs=[method3, hybrid_method3], outputs=[hybrid_method3, hop_length3])
-            with gr.Row():
-                hybrid_method3.change(fn=hoplength_show, inputs=[method3, hybrid_method3], outputs=[hop_length3])
-                refesh2.click(fn=change_models_choices, inputs=[], outputs=[model_pth2, model_index2])
-                model_pth2.change(fn=get_index, inputs=[model_pth2], outputs=[model_index2])
+                            pretrain_choices = gr.Dropdown(label=translations["select_pretrain"], info=translations["select_pretrain_info"], choices=list(fetch_pretrained_data().keys()), value="Titan_Medium", allow_custom_value=True, interactive=True, scale=6, visible=False)
+                            sample_rate_pretrain = gr.Dropdown(label=translations["pretrain_sr"], info=translations["pretrain_sr"], choices=["48k", "40k", "32k"], value="48k", interactive=True, visible=False)
+                        download_pretrain_choices_button = gr.Button(translations["downloads"], scale=2, variant="primary", visible=False)
+                    with gr.Row():
+                        pretrain_upload_g = gr.File(label=translations["drop_pretrain"].format(dg="G"), file_types=[".pth"], visible=False)
+                        pretrain_upload_d = gr.File(label=translations["drop_pretrain"].format(dg="D"), file_types=[".pth"], visible=False)
             with gr.Row():
-                refesh3.click(fn=change_models_choices, inputs=[], outputs=[model_pth3, model_index3])
-                model_pth3.change(fn=get_index, inputs=[model_pth3], outputs=[model_index3])
-                input2.upload(fn=lambda audio_in: shutil.move(audio_in.name, os.path.join("audios")), inputs=[input2], outputs=[input_audio1])
+                url_download.click(
+                    fn=download_model, 
+                    inputs=[
+                        url_input, 
+                        download_model_name
+                    ], 
+                    outputs=[url_input],
+                    api_name="download_model"
+                )
+                download_from_browser.click(
+                    fn=lambda model: download_model(models[model], model), 
+                    inputs=[model_browser], 
+                    outputs=[model_browser],
+                    api_name="download_browser"
+                )
             with gr.Row():
-                input_audio1.change(fn=lambda audio: audio if os.path.isfile(audio) else None, inputs=[input_audio1], outputs=[play_audio2])
-                formant_shifting2.change(fn=lambda a: [visible(a)]*4, inputs=[formant_shifting2], outputs=[formant_qfrency3, formant_timbre3, formant_qfrency4, formant_timbre4])
-                embedders3.change(fn=lambda embedders: visible(embedders == "custom"), inputs=[embedders3], outputs=[custom_embedders3])
+                downloadmodel.change(fn=change_download_choices, inputs=[downloadmodel], outputs=[url_input, download_model_name, url_download, model_browser, download_from_browser, search_name, search, search_dropdown, download, model_upload])
+                search.click(fn=search_models, inputs=[search_name], outputs=[search_dropdown, download])
+                model_upload.upload(fn=save_drop_model, inputs=[model_upload], outputs=[model_upload])
+                download.click(
+                    fn=lambda model: download_model(model_options[model], model), 
+                    inputs=[search_dropdown], 
+                    outputs=[search_dropdown],
+                    api_name="search_models"
+                )
             with gr.Row():
-                refesh4.click(fn=change_audios_choices, inputs=[input_audio1], outputs=[input_audio1])
-                model_index2.change(fn=index_strength_show, inputs=[model_index2], outputs=[index_strength2])
-                model_index3.change(fn=index_strength_show, inputs=[model_index3], outputs=[index_strength3])
+                pretrain_download_choices.change(fn=change_download_pretrained_choices, inputs=[pretrain_download_choices], outputs=[pretrainD, pretrainG, download_pretrain_button, pretrain_choices, sample_rate_pretrain, download_pretrain_choices_button, pretrain_upload_d, pretrain_upload_g])
+                pretrain_choices.change(fn=update_sample_rate_dropdown, inputs=[pretrain_choices], outputs=[sample_rate_pretrain])
             with gr.Row():
-                unlock_full_method2.change(fn=unlock_f0, inputs=[unlock_full_method2], outputs=[method3])
-                embed_mode3.change(fn=visible_embedders, inputs=[embed_mode3], outputs=[embedders3])
-                convert_button3.click(
-                    fn=convert_with_whisper,
+                download_pretrain_button.click(
+                    fn=download_pretrained_model,
                     inputs=[
-                        num_spk,
-                        model_size,
-                        cleaner2,
-                        clean_strength3,
-                        autotune2,
-                        f0_autotune_strength3,
-                        checkpointing2,
-                        model_pth2,
-                        model_pth3,
-                        model_index2,
-                        model_index3,
-                        pitch3,
-                        pitch4,
-                        index_strength2,
-                        index_strength3,
-                        export_format2,
-                        input_audio1,
-                        output_audio2,
-                        onnx_f0_mode4,
-                        method3,
-                        hybrid_method3,
-                        hop_length3,
-                        embed_mode3,
-                        embedders3,
-                        custom_embedders3,
-                        resample_sr3,
-                        filter_radius3,
-                        volume_envelope3,
-                        protect3,
-                        formant_shifting2,
-                        formant_qfrency3,
-                        formant_timbre3,
-                        formant_qfrency4,
-                        formant_timbre4,
+                        pretrain_download_choices, 
+                        pretrainD, 
+                        pretrainG
                     ],
-                    outputs=[play_audio3],
-                    api_name="convert_with_whisper"
+                    outputs=[pretrainD],
+                    api_name="download_pretrain_link"
+                )
+                download_pretrain_choices_button.click(
+                    fn=download_pretrained_model,
+                    inputs=[
+                        pretrain_download_choices, 
+                        pretrain_choices, 
+                        sample_rate_pretrain
+                    ],
+                    outputs=[pretrain_choices],
+                    api_name="download_pretrain_choices"
+                )
+                pretrain_upload_g.upload(
+                    fn=lambda pretrain_upload_g: shutil.move(pretrain_upload_g.name, os.path.join("assets", "models", "pretrained_custom")), 
+                    inputs=[pretrain_upload_g], 
+                    outputs=[],
+                    api_name="upload_pretrain_g"
+                )
+                pretrain_upload_d.upload(
+                    fn=lambda pretrain_upload_d: shutil.move(pretrain_upload_d.name, os.path.join("assets", "models", "pretrained_custom")), 
+                    inputs=[pretrain_upload_d], 
+                    outputs=[],
+                    api_name="upload_pretrain_d"
                 )
 
         with gr.TabItem(translations["convert_text"], visible=configs.get("tts_tab", True)):
@@ -723,113 +688,7 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
                     api_name="convert_tts"
                 )
 
-        with gr.TabItem(translations["downloads"], visible=configs.get("downloads_tab", True)):
-            gr.Markdown(translations["download_markdown"])
-            with gr.Row():
-                gr.Markdown(translations["download_markdown_2"])
-            with gr.Row():
-                with gr.Accordion(translations["model_download"], open=True):
-                    with gr.Row():
-                        downloadmodel = gr.Radio(label=translations["model_download_select"], choices=[translations["download_url"], translations["download_from_csv"], translations["search_models"], translations["upload"]], interactive=True, value=translations["download_url"])
-                    with gr.Row():
-                        gr.Markdown("___")
-                    with gr.Column():
-                        with gr.Row():
-                            url_input = gr.Textbox(label=translations["model_url"], value="", placeholder="https://...", scale=6)
-                            download_model_name = gr.Textbox(label=translations["modelname"], value="", placeholder=translations["modelname"], scale=2)
-                        url_download = gr.Button(value=translations["downloads"], scale=2)
-                    with gr.Column():
-                        model_browser = gr.Dropdown(choices=models.keys(), label=translations["model_warehouse"], scale=8, allow_custom_value=True, visible=False)
-                        download_from_browser = gr.Button(value=translations["get_model"], scale=2, variant="primary", visible=False)
-                    with gr.Column():
-                        search_name = gr.Textbox(label=translations["name_to_search"], placeholder=translations["modelname"], interactive=True, scale=8, visible=False)
-                        search = gr.Button(translations["search_2"], scale=2, visible=False)
-                        search_dropdown = gr.Dropdown(label=translations["select_download_model"], value="", choices=[], allow_custom_value=True, interactive=False, visible=False)
-                        download = gr.Button(translations["downloads"], variant="primary", visible=False)
-                    with gr.Column():
-                        model_upload = gr.File(label=translations["drop_model"], file_types=[".pth", ".onnx", ".index", ".zip"], visible=False)
-            with gr.Row():
-                with gr.Accordion(translations["download_pretrained_2"], open=False):
-                    with gr.Row():
-                        pretrain_download_choices = gr.Radio(label=translations["model_download_select"], choices=[translations["download_url"], translations["list_model"], translations["upload"]], value=translations["download_url"], interactive=True)  
-                    with gr.Row():
-                        gr.Markdown("___")
-                    with gr.Column():
-                        with gr.Row():
-                            pretrainD = gr.Textbox(label=translations["pretrained_url"].format(dg="D"), value="", info=translations["only_huggingface"], placeholder="https://...", interactive=True, scale=4)
-                            pretrainG = gr.Textbox(label=translations["pretrained_url"].format(dg="G"), value="", info=translations["only_huggingface"], placeholder="https://...", interactive=True, scale=4)
-                        download_pretrain_button = gr.Button(translations["downloads"], scale=2)
-                    with gr.Column():
-                        with gr.Row():
-                            pretrain_choices = gr.Dropdown(label=translations["select_pretrain"], info=translations["select_pretrain_info"], choices=list(fetch_pretrained_data().keys()), value="Titan_Medium", allow_custom_value=True, interactive=True, scale=6, visible=False)
-                            sample_rate_pretrain = gr.Dropdown(label=translations["pretrain_sr"], info=translations["pretrain_sr"], choices=["48k", "40k", "32k"], value="48k", interactive=True, visible=False)
-                        download_pretrain_choices_button = gr.Button(translations["downloads"], scale=2, variant="primary", visible=False)
-                    with gr.Row():
-                        pretrain_upload_g = gr.File(label=translations["drop_pretrain"].format(dg="G"), file_types=[".pth"], visible=False)
-                        pretrain_upload_d = gr.File(label=translations["drop_pretrain"].format(dg="D"), file_types=[".pth"], visible=False)
-            with gr.Row():
-                url_download.click(
-                    fn=download_model, 
-                    inputs=[
-                        url_input, 
-                        download_model_name
-                    ], 
-                    outputs=[url_input],
-                    api_name="download_model"
-                )
-                download_from_browser.click(
-                    fn=lambda model: download_model(models[model], model), 
-                    inputs=[model_browser], 
-                    outputs=[model_browser],
-                    api_name="download_browser"
-                )
-            with gr.Row():
-                downloadmodel.change(fn=change_download_choices, inputs=[downloadmodel], outputs=[url_input, download_model_name, url_download, model_browser, download_from_browser, search_name, search, search_dropdown, download, model_upload])
-                search.click(fn=search_models, inputs=[search_name], outputs=[search_dropdown, download])
-                model_upload.upload(fn=save_drop_model, inputs=[model_upload], outputs=[model_upload])
-                download.click(
-                    fn=lambda model: download_model(model_options[model], model), 
-                    inputs=[search_dropdown], 
-                    outputs=[search_dropdown],
-                    api_name="search_models"
-                )
-            with gr.Row():
-                pretrain_download_choices.change(fn=change_download_pretrained_choices, inputs=[pretrain_download_choices], outputs=[pretrainD, pretrainG, download_pretrain_button, pretrain_choices, sample_rate_pretrain, download_pretrain_choices_button, pretrain_upload_d, pretrain_upload_g])
-                pretrain_choices.change(fn=update_sample_rate_dropdown, inputs=[pretrain_choices], outputs=[sample_rate_pretrain])
-            with gr.Row():
-                download_pretrain_button.click(
-                    fn=download_pretrained_model,
-                    inputs=[
-                        pretrain_download_choices, 
-                        pretrainD, 
-                        pretrainG
-                    ],
-                    outputs=[pretrainD],
-                    api_name="download_pretrain_link"
-                )
-                download_pretrain_choices_button.click(
-                    fn=download_pretrained_model,
-                    inputs=[
-                        pretrain_download_choices, 
-                        pretrain_choices, 
-                        sample_rate_pretrain
-                    ],
-                    outputs=[pretrain_choices],
-                    api_name="download_pretrain_choices"
-                )
-                pretrain_upload_g.upload(
-                    fn=lambda pretrain_upload_g: shutil.move(pretrain_upload_g.name, os.path.join("assets", "models", "pretrained_custom")), 
-                    inputs=[pretrain_upload_g], 
-                    outputs=[],
-                    api_name="upload_pretrain_g"
-                )
-                pretrain_upload_d.upload(
-                    fn=lambda pretrain_upload_d: shutil.move(pretrain_upload_d.name, os.path.join("assets", "models", "pretrained_custom")), 
-                    inputs=[pretrain_upload_d], 
-                    outputs=[],
-                    api_name="upload_pretrain_d"
-                )
-
+        
         with gr.TabItem(translations["training_model"], visible=configs.get("training_tab", True)):
             gr.Markdown(f"## {translations['training_model']}")
             with gr.Row():
@@ -1041,6 +900,152 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
                     api_name="training_model"
                 )
 
+        with gr.TabItem(translations["convert_with_whisper"], visible=configs.get("convert_with_whisper", True)):
+            gr.Markdown(f"## {translations['convert_with_whisper']}")
+            with gr.Row():
+                gr.Markdown(translations["convert_with_whisper_info"])
+            with gr.Row():
+                with gr.Column():
+                    with gr.Group():
+                        with gr.Row():
+                            cleaner2 = gr.Checkbox(label=translations["clear_audio"], value=False, interactive=True)
+                            autotune2 = gr.Checkbox(label=translations["autotune"], value=False, interactive=True)
+                            checkpointing2 = gr.Checkbox(label=translations["memory_efficient_training"], value=False, interactive=True)
+                            formant_shifting2 = gr.Checkbox(label=translations["formantshift"], value=False, interactive=True)
+                        with gr.Row():
+                            num_spk = gr.Slider(minimum=2, maximum=8, step=1, info=translations["num_spk_info"], label=translations["num_spk"], value=2, interactive=True)
+            with gr.Row():
+                with gr.Column():
+                    convert_button3 = gr.Button(translations["convert_audio"], variant="primary")
+            with gr.Row():
+                with gr.Column():
+                    with gr.Accordion(translations["model_accordion"] + " 1", open=True):
+                        with gr.Row():
+                            model_pth2 = gr.Dropdown(label=translations["model_name"], choices=model_name, value=model_name[0] if len(model_name) >= 1 else "", interactive=True, allow_custom_value=True)
+                            model_index2 = gr.Dropdown(label=translations["index_path"], choices=index_path, value=index_path[0] if len(index_path) >= 1 else "", interactive=True, allow_custom_value=True)
+                        with gr.Row():
+                            refesh2 = gr.Button(translations["refesh"])
+                        with gr.Row():
+                            pitch3 = gr.Slider(minimum=-20, maximum=20, step=1, info=translations["pitch_info"], label=translations["pitch"], value=0, interactive=True)
+                            index_strength2 = gr.Slider(label=translations["index_strength"], info=translations["index_strength_info"], minimum=0, maximum=1, value=0.5, step=0.01, interactive=True, visible=model_index2.value != "")
+                    with gr.Accordion(translations["input_output"], open=False):
+                        with gr.Column():
+                            export_format2 = gr.Radio(label=translations["export_format"], info=translations["export_info"], choices=["wav", "mp3", "flac", "ogg", "opus", "m4a", "mp4", "aac", "alac", "wma", "aiff", "webm", "ac3"], value="wav", interactive=True)
+                            input_audio1 = gr.Dropdown(label=translations["audio_path"], value="", choices=paths_for_files, info=translations["provide_audio"], allow_custom_value=True, interactive=True)
+                            output_audio2 = gr.Textbox(label=translations["output_path"], value="audios/output.wav", placeholder="audios/output.wav", info=translations["output_path_info"], interactive=True)
+                        with gr.Column():
+                            refesh4 = gr.Button(translations["refesh"])
+                        with gr.Row():
+                            input2 = gr.File(label=translations["drop_audio"], file_types=[".wav", ".mp3", ".flac", ".ogg", ".opus", ".m4a", ".mp4", ".aac", ".alac", ".wma", ".aiff", ".webm", ".ac3"])
+                with gr.Column():
+                    with gr.Accordion(translations["model_accordion"] + " 2", open=True):
+                        with gr.Row():
+                            model_pth3 = gr.Dropdown(label=translations["model_name"], choices=model_name, value=model_name[0] if len(model_name) >= 1 else "", interactive=True, allow_custom_value=True)
+                            model_index3 = gr.Dropdown(label=translations["index_path"], choices=index_path, value=index_path[0] if len(index_path) >= 1 else "", interactive=True, allow_custom_value=True)
+                        with gr.Row():
+                            refesh3 = gr.Button(translations["refesh"])
+                        with gr.Row():
+                            pitch4 = gr.Slider(minimum=-20, maximum=20, step=1, info=translations["pitch_info"], label=translations["pitch"], value=0, interactive=True)
+                            index_strength3 = gr.Slider(label=translations["index_strength"], info=translations["index_strength_info"], minimum=0, maximum=1, value=0.5, step=0.01, interactive=True, visible=model_index3.value != "")
+                    with gr.Accordion(translations["setting"], open=False):
+                        with gr.Row():
+                            model_size = gr.Radio(label=translations["model_size"], info=translations["model_size_info"], choices=["tiny", "tiny.en", "base", "base.en", "small", "small.en", "medium", "medium.en", "large-v1", "large-v2", "large-v3", "large-v3-turbo"], value="medium", interactive=True)
+                        with gr.Accordion(translations["f0_method"], open=False):
+                            with gr.Group():
+                                with gr.Row():
+                                    onnx_f0_mode4 = gr.Checkbox(label=translations["f0_onnx_mode"], info=translations["f0_onnx_mode_info"], value=False, interactive=True)
+                                    unlock_full_method2 = gr.Checkbox(label=translations["f0_unlock"], info=translations["f0_unlock_info"], value=False, interactive=True)
+                                method3 = gr.Radio(label=translations["f0_method"], info=translations["f0_method_info"], choices=method_f0+["hybrid"], value="rmvpe", interactive=True)
+                                hybrid_method3 = gr.Dropdown(label=translations["f0_method_hybrid"], info=translations["f0_method_hybrid_info"], choices=["hybrid[pm+dio]", "hybrid[pm+crepe-tiny]", "hybrid[pm+crepe]", "hybrid[pm+fcpe]", "hybrid[pm+rmvpe]", "hybrid[pm+harvest]", "hybrid[pm+yin]", "hybrid[dio+crepe-tiny]", "hybrid[dio+crepe]", "hybrid[dio+fcpe]", "hybrid[dio+rmvpe]", "hybrid[dio+harvest]", "hybrid[dio+yin]", "hybrid[crepe-tiny+crepe]", "hybrid[crepe-tiny+fcpe]", "hybrid[crepe-tiny+rmvpe]", "hybrid[crepe-tiny+harvest]", "hybrid[crepe+fcpe]", "hybrid[crepe+rmvpe]", "hybrid[crepe+harvest]", "hybrid[crepe+yin]", "hybrid[fcpe+rmvpe]", "hybrid[fcpe+harvest]", "hybrid[fcpe+yin]", "hybrid[rmvpe+harvest]", "hybrid[rmvpe+yin]", "hybrid[harvest+yin]"], value="hybrid[pm+dio]", interactive=True, allow_custom_value=True, visible=method3.value == "hybrid")
+                            hop_length3 = gr.Slider(label="Hop length", info=translations["hop_length_info"], minimum=1, maximum=512, value=128, step=1, interactive=True, visible=False)
+                        with gr.Accordion(translations["hubert_model"], open=False):
+                            embed_mode3 = gr.Radio(label=translations["embed_mode"], info=translations["embed_mode_info"], value="fairseq", choices=embedders_mode, interactive=True, visible=True)
+                            embedders3 = gr.Radio(label=translations["hubert_model"], info=translations["hubert_info"], choices=embedders_model, value="hubert_base", interactive=True)
+                            custom_embedders3 = gr.Textbox(label=translations["modelname"], info=translations["modelname_info"], value="", placeholder="hubert_base", interactive=True, visible=embedders3.value == "custom")
+                        with gr.Column():      
+                            clean_strength3 = gr.Slider(label=translations["clean_strength"], info=translations["clean_strength_info"], minimum=0, maximum=1, value=0.5, step=0.1, interactive=True, visible=cleaner2.value)
+                            f0_autotune_strength3 = gr.Slider(minimum=0, maximum=1, label=translations["autotune_rate"], info=translations["autotune_rate_info"], value=1, step=0.1, interactive=True, visible=autotune.value)
+                            resample_sr3 = gr.Slider(minimum=0, maximum=96000, label=translations["resample"], info=translations["resample_info"], value=0, step=1, interactive=True)
+                            filter_radius3 = gr.Slider(minimum=0, maximum=7, label=translations["filter_radius"], info=translations["filter_radius_info"], value=3, step=1, interactive=True)
+                            volume_envelope3 = gr.Slider(minimum=0, maximum=1, label=translations["volume_envelope"], info=translations["volume_envelope_info"], value=1, step=0.1, interactive=True)
+                            protect3 = gr.Slider(minimum=0, maximum=1, label=translations["protect"], info=translations["protect_info"], value=0.5, step=0.01, interactive=True)
+                        with gr.Row():
+                            formant_qfrency3 = gr.Slider(value=1.0, label=translations["formant_qfrency"] + " 1", info=translations["formant_qfrency"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
+                            formant_timbre3 = gr.Slider(value=1.0, label=translations["formant_timbre"] + " 1", info=translations["formant_timbre"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
+                        with gr.Row():
+                            formant_qfrency4 = gr.Slider(value=1.0, label=translations["formant_qfrency"] + " 2", info=translations["formant_qfrency"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
+                            formant_timbre4 = gr.Slider(value=1.0, label=translations["formant_timbre"] + " 2", info=translations["formant_timbre"], minimum=0.0, maximum=16.0, step=0.1, interactive=True, visible=False)
+            with gr.Row():
+                gr.Markdown(translations["input_output"])
+            with gr.Row():
+                play_audio2 = gr.Audio(show_download_button=True, interactive=False, label=translations["input_audio"])
+                play_audio3 = gr.Audio(show_download_button=True, interactive=False, label=translations["output_file_tts_convert"])
+            with gr.Row():
+                autotune2.change(fn=visible, inputs=[autotune2], outputs=[f0_autotune_strength3])
+                cleaner2.change(fn=visible, inputs=[cleaner2], outputs=[clean_strength3])
+                method3.change(fn=lambda method, hybrid: [visible(method == "hybrid"), hoplength_show(method, hybrid)], inputs=[method3, hybrid_method3], outputs=[hybrid_method3, hop_length3])
+            with gr.Row():
+                hybrid_method3.change(fn=hoplength_show, inputs=[method3, hybrid_method3], outputs=[hop_length3])
+                refesh2.click(fn=change_models_choices, inputs=[], outputs=[model_pth2, model_index2])
+                model_pth2.change(fn=get_index, inputs=[model_pth2], outputs=[model_index2])
+            with gr.Row():
+                refesh3.click(fn=change_models_choices, inputs=[], outputs=[model_pth3, model_index3])
+                model_pth3.change(fn=get_index, inputs=[model_pth3], outputs=[model_index3])
+                input2.upload(fn=lambda audio_in: shutil.move(audio_in.name, os.path.join("audios")), inputs=[input2], outputs=[input_audio1])
+            with gr.Row():
+                input_audio1.change(fn=lambda audio: audio if os.path.isfile(audio) else None, inputs=[input_audio1], outputs=[play_audio2])
+                formant_shifting2.change(fn=lambda a: [visible(a)]*4, inputs=[formant_shifting2], outputs=[formant_qfrency3, formant_timbre3, formant_qfrency4, formant_timbre4])
+                embedders3.change(fn=lambda embedders: visible(embedders == "custom"), inputs=[embedders3], outputs=[custom_embedders3])
+            with gr.Row():
+                refesh4.click(fn=change_audios_choices, inputs=[input_audio1], outputs=[input_audio1])
+                model_index2.change(fn=index_strength_show, inputs=[model_index2], outputs=[index_strength2])
+                model_index3.change(fn=index_strength_show, inputs=[model_index3], outputs=[index_strength3])
+            with gr.Row():
+                unlock_full_method2.change(fn=unlock_f0, inputs=[unlock_full_method2], outputs=[method3])
+                embed_mode3.change(fn=visible_embedders, inputs=[embed_mode3], outputs=[embedders3])
+                convert_button3.click(
+                    fn=convert_with_whisper,
+                    inputs=[
+                        num_spk,
+                        model_size,
+                        cleaner2,
+                        clean_strength3,
+                        autotune2,
+                        f0_autotune_strength3,
+                        checkpointing2,
+                        model_pth2,
+                        model_pth3,
+                        model_index2,
+                        model_index3,
+                        pitch3,
+                        pitch4,
+                        index_strength2,
+                        index_strength3,
+                        export_format2,
+                        input_audio1,
+                        output_audio2,
+                        onnx_f0_mode4,
+                        method3,
+                        hybrid_method3,
+                        hop_length3,
+                        embed_mode3,
+                        embedders3,
+                        custom_embedders3,
+                        resample_sr3,
+                        filter_radius3,
+                        volume_envelope3,
+                        protect3,
+                        formant_shifting2,
+                        formant_qfrency3,
+                        formant_timbre3,
+                        formant_qfrency4,
+                        formant_timbre4,
+                    ],
+                    outputs=[play_audio3],
+                    api_name="convert_with_whisper"
+                )
+
+        
         with gr.TabItem(translations["audio_editing"], visible=configs.get("audioldm2", True)):
             gr.Markdown(translations["audio_editing_info"])
             with gr.Row():
@@ -1537,7 +1542,7 @@ with gr.Blocks(title=" Ultimate RVC Maker ⚡", theme=theme) as app:
     
     with gr.Row(): 
         gr.Markdown(translations["terms_of_use"])
-    with gr.Row():
+    
         gr.Markdown(translations["exemption"])
 
     logger.info(translations["start_app"])