Hev832 commited on
Commit
a5cea98
·
verified ·
1 Parent(s): 69d13ea

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +97 -32
app.py CHANGED
@@ -95,48 +95,94 @@ if __name__ == '__main__':
95
 
96
  # Gradio Interface
97
  with gr.Blocks(title="Hex RVC", theme=gr.themes.Base(primary_hue="red", secondary_hue="pink")) as app:
98
- gr.Markdown("# Hex RVC")
99
  gr.Markdown("Join [AIHub](https://discord.gg/aihub) to get the RVC model!")
100
-
 
101
  with gr.Tab("Inference"):
 
102
  with gr.Row():
103
  MODEL_NAME = gr.Dropdown(
104
- label="Select a Model",
105
  choices=get_folders(),
106
- interactive=True
 
107
  )
108
  SOUND_PATH = gr.Dropdown(
109
  choices=load_audio_files(),
110
- label="Select an audio file",
111
- interactive=True
 
112
  )
113
- upload_audio = gr.Audio(label="Upload Audio", type='filepath')
114
-
115
- with gr.Accordion("Hex TTS"):
116
- input_text = gr.Textbox(lines=5, label="Input Text")
117
- language = gr.Dropdown(choices=list(language_dict.keys()), label="Choose the Voice Model")
118
- tts_convert = gr.Button("Convert")
119
- tts_output = gr.Audio(label="Generated TTS Audio", type='filepath')
120
-
121
- tts_convert.click(
122
- fn=text_to_speech_edge,
123
- inputs=[input_text, language],
124
- outputs=tts_output
125
  )
126
 
127
- output_audio = gr.Audio(label="Generated Audio", type='filepath')
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
128
  with gr.Row():
129
- refresh_btn = gr.Button("Refresh")
130
- run_button = gr.Button("Convert")
131
 
 
132
  refresh_btn.click(
133
  lambda: (refresh_audio_list(), refresh_folders()),
134
  outputs=[SOUND_PATH, MODEL_NAME]
135
  )
 
 
 
 
 
 
 
 
 
136
 
 
137
  with gr.Tab("Download RVC Model"):
138
- url = gr.Textbox(label="Your Model URL")
139
- dirname = gr.Textbox(label="Your Model Name")
 
140
  download_button = gr.Button("Download Model")
141
  download_output = gr.Textbox(label="Download Status")
142
 
@@ -147,17 +193,36 @@ with gr.Blocks(title="Hex RVC", theme=gr.themes.Base(primary_hue="red", secondar
147
  )
148
 
149
  with gr.Tab("Audio Separation"):
150
- input_audio = gr.Audio(type="filepath", label="Upload Audio")
151
- roformer_link = gr.Textbox(label="Audio Link")
152
- roformer_download_button = gr.Button("Download")
153
-
154
- separate_button = gr.Button("Separate Audio")
155
- separation_output = gr.Textbox(label="Separation Output Path")
156
-
157
- roformer_download_button.click(download_audio, inputs=[roformer_link], outputs=[input_audio])
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
158
 
159
  app.launch(
160
  share=args.share_enabled,
161
  server_name=None if not args.listen else (args.listen_host or '0.0.0.0'),
162
- server_port=args.listen_port,
163
  )
 
95
 
96
  # Gradio Interface
97
  with gr.Blocks(title="Hex RVC", theme=gr.themes.Base(primary_hue="red", secondary_hue="pink")) as app:
98
+ gr.Markdown("# Hex RVC - AI Audio Inference")
99
  gr.Markdown("Join [AIHub](https://discord.gg/aihub) to get the RVC model!")
100
+
101
+ # Inference Tab with Priority on Settings
102
  with gr.Tab("Inference"):
103
+ gr.Markdown("## Inference Settings")
104
  with gr.Row():
105
  MODEL_NAME = gr.Dropdown(
106
+ label="Select AI Model",
107
  choices=get_folders(),
108
+ interactive=True,
109
+ info="Choose a pre-trained model for audio processing"
110
  )
111
  SOUND_PATH = gr.Dropdown(
112
  choices=load_audio_files(),
113
+ label="Select Existing Audio File",
114
+ interactive=True,
115
+ info="Pick an audio file from the predefined directory"
116
  )
117
+ upload_audio = gr.Audio(
118
+ label="Upload Your Own Audio",
119
+ type='filepath',
120
+ info="Upload an audio file if not using existing ones"
 
 
 
 
 
 
 
 
121
  )
122
 
123
+ gr.Markdown("### Conversion Parameters")
124
+ with gr.Accordion("Conversion Settings", open=True):
125
+ with gr.Row():
126
+ F0_CHANGE = gr.Number(
127
+ label="Pitch Change (semitones)",
128
+ value=0,
129
+ info="Adjust the pitch of the output audio"
130
+ )
131
+ F0_METHOD = gr.Dropdown(
132
+ choices=["crepe", "harvest", "mangio-crepe", "rmvpe", "rmvpe_legacy", "fcpe", "fcpe_legacy", "hybrid[rmvpe+fcpe]"],
133
+ label="F0 Method",
134
+ value="fcpe",
135
+ info="Select the fundamental frequency extraction method"
136
+ )
137
+ with gr.Row():
138
+ MIN_PITCH = gr.Number(label="Min Pitch", value=50, info="Minimum pitch detection threshold")
139
+ MAX_PITCH = gr.Number(label="Max Pitch", value=1100, info="Maximum pitch detection threshold")
140
+ CREPE_HOP_LENGTH = gr.Number(label="Crepe Hop Length", value=120, info="Hop length for Crepe method")
141
+ INDEX_RATE = gr.Slider(label="Index Rate", minimum=0, maximum=1, value=0.75)
142
+ FILTER_RADIUS = gr.Number(label="Filter Radius", value=3, info="Filter intensity for smoothing")
143
+ RMS_MIX_RATE = gr.Slider(label="RMS Mix Rate", minimum=0, maximum=1, value=0.25)
144
+ PROTECT = gr.Slider(label="Protect Factor", minimum=0, maximum=1, value=0.33)
145
+
146
+ gr.Markdown("### Advanced Settings")
147
+ with gr.Accordion("Advanced Settings", open=False):
148
+ SPLIT_INFER = gr.Checkbox(label="Enable Split Inference", value=False)
149
+ MIN_SILENCE = gr.Number(label="Min Silence (ms)", value=500)
150
+ SILENCE_THRESHOLD = gr.Number(label="Silence Threshold (dBFS)", value=-50)
151
+ SEEK_STEP = gr.Slider(label="Seek Step (ms)", minimum=1, maximum=10, value=1)
152
+ KEEP_SILENCE = gr.Number(label="Keep Silence (ms)", value=200)
153
+ FORMANT_SHIFT = gr.Checkbox(label="Enable Formant Shift", value=False)
154
+ QUEFRENCY = gr.Number(label="Quefrency", value=0)
155
+ TIMBRE = gr.Number(label="Timbre", value=1)
156
+ F0_AUTOTUNE = gr.Checkbox(label="Enable F0 Autotune", value=False)
157
+ OUTPUT_FORMAT = gr.Dropdown(choices=["wav", "flac", "mp3"], label="Output Format", value="wav")
158
+
159
+ gr.Markdown("## Generate Audio")
160
+ output_audio = gr.Audio(label="Generated Audio Output", type='filepath')
161
+
162
  with gr.Row():
163
+ refresh_btn = gr.Button("Refresh Lists")
164
+ run_button = gr.Button("Run Inference")
165
 
166
+ # Refresh Button for Updating Model and Audio Choices
167
  refresh_btn.click(
168
  lambda: (refresh_audio_list(), refresh_folders()),
169
  outputs=[SOUND_PATH, MODEL_NAME]
170
  )
171
+
172
+ # Run Inference and Display Result
173
+ run_button.click(
174
+ fn=process_audio,
175
+ inputs=[MODEL_NAME, SOUND_PATH, F0_CHANGE, F0_METHOD, MIN_PITCH, MAX_PITCH, CREPE_HOP_LENGTH, INDEX_RATE,
176
+ FILTER_RADIUS, RMS_MIX_RATE, PROTECT, SPLIT_INFER, MIN_SILENCE, SILENCE_THRESHOLD, SEEK_STEP,
177
+ KEEP_SILENCE, FORMANT_SHIFT, QUEFRENCY, TIMBRE, F0_AUTOTUNE, OUTPUT_FORMAT, upload_audio],
178
+ outputs=output_audio
179
+ )
180
 
181
+ # Other Tabs (Download Model, Audio Separation)
182
  with gr.Tab("Download RVC Model"):
183
+ gr.Markdown("## Download RVC Model")
184
+ url = gr.Textbox(label="Model URL")
185
+ dirname = gr.Textbox(label="Model Directory Name")
186
  download_button = gr.Button("Download Model")
187
  download_output = gr.Textbox(label="Download Status")
188
 
 
193
  )
194
 
195
  with gr.Tab("Audio Separation"):
196
+ gr.Markdown("## Audio Separation")
197
+ input_audio = gr.Audio(type="filepath", label="Upload Audio for Separation")
198
+ with gr.Accordion("Separation by Link", open = False):
199
+ with gr.Row():
200
+ roformer_link = gr.Textbox(
201
+ label = "Link",
202
+ placeholder = "Paste the link here",
203
+ interactive = True
204
+ )
205
+ with gr.Row():
206
+ gr.Markdown("You can paste the link to the video/audio from many sites, check the complete list [here](https://github.com/yt-dlp/yt-dlp/blob/master/supportedsites.md)")
207
+ with gr.Row():
208
+ roformer_download_button = gr.Button(
209
+ "Download!",
210
+ variant = "primary"
211
+ )
212
+ separate_button = gr.Button("Separate Audio")
213
+ separation_output = gr.Textbox(label="Separation Output Path")
214
+
215
+ roformer_download_button.click(download_audio, [roformer_link], [input_audio])
216
+ separate_button.click(
217
+ fn=separate_audio,
218
+ inputs=[input_audio, "model_bs_roformer_ep_317_sdr_12.9755.ckpt",
219
+ "UVR-DeEcho-DeReverb.pth",
220
+ "mel_band_roformer_karaoke_aufr33_viperx_sdr_10.1956.ckpt"],
221
+ outputs=[separation_output]
222
+ )
223
 
224
  app.launch(
225
  share=args.share_enabled,
226
  server_name=None if not args.listen else (args.listen_host or '0.0.0.0'),
227
+ server_port=args.listen_port
228
  )