Update app.py

app.py

@@ -30,104 +30,104 @@ import psutil
 whisper_models = ["tiny", "base", "small", "medium", "large-v1", "large-v2"]
 source_languages = {
     "en": "English",
-    # "zh": "Chinese",
-    # "de": "German",
-    # "es": "Spanish",
-    # "ru": "Russian",
-    # "ko": "Korean",
-    # "fr": "French",
+    "zh": "Chinese",
+    "de": "German",
+    "es": "Spanish",
+    "ru": "Russian",
+    "ko": "Korean",
+    "fr": "French",
     "ja": "Japanese",
-    # "pt": "Portuguese",
-    # "tr": "Turkish",
-    # "pl": "Polish",
-    # "ca": "Catalan",
-    # "nl": "Dutch",
-    # "ar": "Arabic",
-    # "sv": "Swedish",
-    # "it": "Italian",
-    # "id": "Indonesian",
-    # "hi": "Hindi",
-    # "fi": "Finnish",
-    # "vi": "Vietnamese",
-    # "he": "Hebrew",
-    # "uk": "Ukrainian",
-    # "el": "Greek",
-    # "ms": "Malay",
-    # "cs": "Czech",
-    # "ro": "Romanian",
-    # "da": "Danish",
-    # "hu": "Hungarian",
-    # "ta": "Tamil",
-    # "no": "Norwegian",
-    # "th": "Thai",
-    # "ur": "Urdu",
-    # "hr": "Croatian",
-    # "bg": "Bulgarian",
-    # "lt": "Lithuanian",
-    # "la": "Latin",
-    # "mi": "Maori",
-    # "ml": "Malayalam",
-    # "cy": "Welsh",
-    # "sk": "Slovak",
-    # "te": "Telugu",
-    # "fa": "Persian",
-    # "lv": "Latvian",
-    # "bn": "Bengali",
-    # "sr": "Serbian",
-    # "az": "Azerbaijani",
-    # "sl": "Slovenian",
-    # "kn": "Kannada",
-    # "et": "Estonian",
-    # "mk": "Macedonian",
-    # "br": "Breton",
-    # "eu": "Basque",
-    # "is": "Icelandic",
-    # "hy": "Armenian",
-    # "ne": "Nepali",
-    # "mn": "Mongolian",
-    # "bs": "Bosnian",
-    # "kk": "Kazakh",
-    # "sq": "Albanian",
-    # "sw": "Swahili",
-    # "gl": "Galician",
-    # "mr": "Marathi",
-    # "pa": "Punjabi",
-    # "si": "Sinhala",
-    # "km": "Khmer",
-    # "sn": "Shona",
-    # "yo": "Yoruba",
-    # "so": "Somali",
-    # "af": "Afrikaans",
-    # "oc": "Occitan",
-    # "ka": "Georgian",
-    # "be": "Belarusian",
-    # "tg": "Tajik",
-    # "sd": "Sindhi",
-    # "gu": "Gujarati",
-    # "am": "Amharic",
-    # "yi": "Yiddish",
-    # "lo": "Lao",
-    # "uz": "Uzbek",
-    # "fo": "Faroese",
-    # "ht": "Haitian creole",
-    # "ps": "Pashto",
-    # "tk": "Turkmen",
-    # "nn": "Nynorsk",
-    # "mt": "Maltese",
-    # "sa": "Sanskrit",
-    # "lb": "Luxembourgish",
-    # "my": "Myanmar",
-    # "bo": "Tibetan",
-    # "tl": "Tagalog",
-    # "mg": "Malagasy",
-    # "as": "Assamese",
-    # "tt": "Tatar",
-    # "haw": "Hawaiian",
-    # "ln": "Lingala",
-    # "ha": "Hausa",
-    # "ba": "Bashkir",
-    # "jw": "Javanese",
-    # "su": "Sundanese",
+    "pt": "Portuguese",
+    "tr": "Turkish",
+    "pl": "Polish",
+    "ca": "Catalan",
+    "nl": "Dutch",
+    "ar": "Arabic",
+    "sv": "Swedish",
+    "it": "Italian",
+    "id": "Indonesian",
+    "hi": "Hindi",
+    "fi": "Finnish",
+    "vi": "Vietnamese",
+    "he": "Hebrew",
+    "uk": "Ukrainian",
+    "el": "Greek",
+    "ms": "Malay",
+    "cs": "Czech",
+    "ro": "Romanian",
+    "da": "Danish",
+    "hu": "Hungarian",
+    "ta": "Tamil",
+    "no": "Norwegian",
+    "th": "Thai",
+    "ur": "Urdu",
+    "hr": "Croatian",
+    "bg": "Bulgarian",
+    "lt": "Lithuanian",
+    "la": "Latin",
+    "mi": "Maori",
+    "ml": "Malayalam",
+    "cy": "Welsh",
+    "sk": "Slovak",
+    "te": "Telugu",
+    "fa": "Persian",
+    "lv": "Latvian",
+    "bn": "Bengali",
+    "sr": "Serbian",
+    "az": "Azerbaijani",
+    "sl": "Slovenian",
+    "kn": "Kannada",
+    "et": "Estonian",
+    "mk": "Macedonian",
+    "br": "Breton",
+    "eu": "Basque",
+    "is": "Icelandic",
+    "hy": "Armenian",
+    "ne": "Nepali",
+    "mn": "Mongolian",
+    "bs": "Bosnian",
+    "kk": "Kazakh",
+    "sq": "Albanian",
+    "sw": "Swahili",
+    "gl": "Galician",
+    "mr": "Marathi",
+    "pa": "Punjabi",
+    "si": "Sinhala",
+    "km": "Khmer",
+    "sn": "Shona",
+    "yo": "Yoruba",
+    "so": "Somali",
+    "af": "Afrikaans",
+    "oc": "Occitan",
+    "ka": "Georgian",
+    "be": "Belarusian",
+    "tg": "Tajik",
+    "sd": "Sindhi",
+    "gu": "Gujarati",
+    "am": "Amharic",
+    "yi": "Yiddish",
+    "lo": "Lao",
+    "uz": "Uzbek",
+    "fo": "Faroese",
+    "ht": "Haitian creole",
+    "ps": "Pashto",
+    "tk": "Turkmen",
+    "nn": "Nynorsk",
+    "mt": "Maltese",
+    "sa": "Sanskrit",
+    "lb": "Luxembourgish",
+    "my": "Myanmar",
+    "bo": "Tibetan",
+    "tl": "Tagalog",
+    "mg": "Malagasy",
+    "as": "Assamese",
+    "tt": "Tatar",
+    "haw": "Hawaiian",
+    "ln": "Lingala",
+    "ha": "Hausa",
+    "ba": "Bashkir",
+    "jw": "Javanese",
+    "su": "Sundanese",
 }
 
 source_language_list = [key[0] for key in source_languages.items()]
@@ -357,7 +357,7 @@ video_in = gr.Video(label="Video file", mirror_webcam=False)
 youtube_url_in = gr.Textbox(label="Youtube url", lines=1, interactive=True)
 df_init = pd.DataFrame(columns=['Start', 'End', 'Speaker', 'Text'])
 memory = psutil.virtual_memory()
-selected_source_lang = gr.Dropdown(choices=source_language_list, type="value", value="ja",
+selected_source_lang = gr.Dropdown(choices=source_language_list, type="value", value="en",
                                    label="Spoken language in video", interactive=True)
 selected_whisper_model = gr.Dropdown(choices=whisper_models, type="value", value="base", label="Selected Whisper model",
                                      interactive=True)
@@ -374,67 +374,107 @@ demo = gr.Blocks(title=title)
 demo.encrypt = False
 
 with demo:
-    # gr.Markdown('''
-    #     <div>
-    #     <h1 style='text-align: center'>Whisper speaker diarization</h1>
-    #     This space uses Whisper models from <a href='https://github.com/openai/whisper' target='_blank'><b>OpenAI</b></a> with <a href='https://github.com/guillaumekln/faster-whisper' target='_blank'><b>CTranslate2</b></a> which is a fast inference engine for Transformer models to recognize the speech (4 times faster than original openai model with same accuracy)
-    #     and ECAPA-TDNN model from <a href='https://github.com/speechbrain/speechbrain' target='_blank'><b>SpeechBrain</b></a> to encode and clasify speakers
-    #     </div>
-    # ''')
-    #
-    # with gr.Row():
-    #     gr.Markdown('''
-    #     ### Transcribe youtube link using OpenAI Whisper
-    #     ##### 1. Using Open AI's Whisper model to seperate audio into segments and generate transcripts.
-    #     ##### 2. Generating speaker embeddings for each segments.
-    #     ##### 3. Applying agglomerative clustering on the embeddings to identify the speaker for each segment.
-    #     ''')
-
-    with gr.Row():
+    with gr.Tab("Whisper speaker diarization"):
         gr.Markdown('''
-            ### You can test by following examples:
+            <div>
+            <h1 style='text-align: center'>Whisper speaker diarization</h1>
+            This space uses Whisper models from <a href='https://github.com/openai/whisper' target='_blank'><b>OpenAI</b></a> with <a href='https://github.com/guillaumekln/faster-whisper' target='_blank'><b>CTranslate2</b></a> which is a fast inference engine for Transformer models to recognize the speech (4 times faster than original openai model with same accuracy)
+            and ECAPA-TDNN model from <a href='https://github.com/speechbrain/speechbrain' target='_blank'><b>SpeechBrain</b></a> to encode and clasify speakers
+            </div>
+        ''')
+
+        with gr.Row():
+            gr.Markdown('''
+            ### Transcribe youtube link using OpenAI Whisper
+            ##### 1. Using Open AI's Whisper model to seperate audio into segments and generate transcripts.
+            ##### 2. Generating speaker embeddings for each segments.
+            ##### 3. Applying agglomerative clustering on the embeddings to identify the speaker for each segment.
             ''')
-        examples = gr.Examples(examples=
-                           ["https://www.youtube.com/watch?v=j7BfEzAFuYc&t=32s",
-                            "https://www.youtube.com/watch?v=-UX0X45sYe4",
-                            "https://www.youtube.com/watch?v=7minSgqi-Gw"],
-                           label="Examples", inputs=[youtube_url_in])
-
-    with gr.Row():
-        with gr.Column():
-            youtube_url_in.render()
-            download_youtube_btn = gr.Button("Download Youtube video")
-            download_youtube_btn.click(get_youtube, [youtube_url_in], [video_in])
-            print(video_in)
-
-    with gr.Row():
-        with gr.Column():
-            video_in.render()
-            with gr.Column():
-                gr.Markdown('''
-                ##### Here you can start the transcription process.
-                ##### Please select the source language for transcription.
-                ##### You can select a range of assumed numbers of speakers.
+
+        with gr.Row():
+            gr.Markdown('''
+                ### You can test by following examples:
                 ''')
-            selected_source_lang.render()
-            selected_whisper_model.render()
-            number_speakers.render()
-            transcribe_btn = gr.Button("Transcribe audio and diarization")
-            transcribe_btn.click(speech_to_text,
-                                 [video_in, selected_source_lang, selected_whisper_model, number_speakers],
-                                 [transcription_df, system_info, download_transcript]
-                                 )
-
-    with gr.Row():
-        gr.Markdown('''
-        ##### Here you will get transcription  output
-        ##### ''')
+        examples = gr.Examples(examples=
+                               ["https://www.youtube.com/watch?v=j7BfEzAFuYc&t=32s",
+                                "https://www.youtube.com/watch?v=-UX0X45sYe4",
+                                "https://www.youtube.com/watch?v=7minSgqi-Gw"],
+                               label="Examples", inputs=[youtube_url_in])
 
-    with gr.Row():
-        with gr.Column():
-            download_transcript.render()
-            transcription_df.render()
-            # system_info.render()
+        with gr.Row():
+            with gr.Column():
+                youtube_url_in.render()
+                download_youtube_btn = gr.Button("Download Youtube video")
+                download_youtube_btn.click(get_youtube, [youtube_url_in], [
+                    video_in])
+                print(video_in)
 
+        with gr.Row():
+            with gr.Column():
+                video_in.render()
+                with gr.Column():
+                    gr.Markdown('''
+                    ##### Here you can start the transcription process.
+                    ##### Please select the source language for transcription.
+                    ##### You can select a range of assumed numbers of speakers.
+                    ''')
+                selected_source_lang.render()
+                selected_whisper_model.render()
+                number_speakers.render()
+                transcribe_btn = gr.Button("Transcribe audio and diarization")
+                transcribe_btn.click(speech_to_text,
+                                     [video_in, selected_source_lang, selected_whisper_model, number_speakers],
+                                     [transcription_df, system_info, download_transcript]
+                                     )
+
+        with gr.Row():
+            gr.Markdown('''
+            ##### Here you will get transcription  output
+            ##### ''')
+
+        with gr.Row():
+            with gr.Column():
+                download_transcript.render()
+                transcription_df.render()
+                system_info.render()
+                gr.Markdown(
+                    '''<center><img src='https://visitor-badge.glitch.me/badge?page_id=WhisperDiarizationSpeakers' alt='visitor badge'><a href="https://opensource.org/licenses/Apache-2.0"><img src='https://img.shields.io/badge/License-Apache_2.0-blue.svg' alt='License: Apache 2.0'></center>''')
+
+    with gr.Tab("Whisper Transcribe Japanese Audio"):
+        gr.Markdown(f'''
+              <div>
+              <h1 style='text-align: center'>Whisper Transcribe Japanese Audio</h1>
+              </div>
+              Transcribe long-form microphone or audio inputs with the click of a button! The fine-tuned
+              checkpoint <a href='https://huggingface.co/{MODEL_NAME}' target='_blank'><b>{MODEL_NAME}</b></a> to transcribe audio files of arbitrary length.
+          ''')
+        microphone = gr.inputs.Audio(source="microphone", type="filepath", optional=True)
+        upload = gr.inputs.Audio(source="upload", type="filepath", optional=True)
+        transcribe_btn = gr.Button("Transcribe Audio")
+        text_output = gr.Textbox()
+        with gr.Row():
+            gr.Markdown('''
+                ### You can test by following examples:
+                ''')
+        examples = gr.Examples(examples=
+                               ["sample1.wav",
+                                "sample2.wav",
+                                ],
+                               label="Examples", inputs=[upload])
+        transcribe_btn.click(transcribe, [microphone, upload], outputs=text_output)
+
+    with gr.Tab("Whisper Transcribe Japanese YouTube"):
+        gr.Markdown(f'''
+              <div>
+              <h1 style='text-align: center'>Whisper Transcribe Japanese YouTube</h1>
+              </div>
+                Transcribe long-form YouTube videos with the click of a button! The fine-tuned checkpoint:
+                <a href='https://huggingface.co/{MODEL_NAME}' target='_blank'><b>{MODEL_NAME}</b></a> to transcribe audio files of arbitrary length.
+            ''')
+        youtube_link = gr.Textbox(label="Youtube url", lines=1, interactive=True)
+        yt_transcribe_btn = gr.Button("Transcribe YouTube")
+        text_output2 = gr.Textbox()
+        html_output = gr.Markdown()
+        yt_transcribe_btn.click(yt_transcribe, [youtube_link], outputs=[html_output, text_output2])
 
 demo.launch(debug=True)