Spaces:
Running
on
Zero
Running
on
Zero
chong.zhang
commited on
Commit
·
d896a67
1
Parent(s):
489e7d6
update
Browse files
app.py
CHANGED
|
@@ -6,8 +6,7 @@ import torchaudio
|
|
| 6 |
import datetime
|
| 7 |
import hashlib
|
| 8 |
import torch
|
| 9 |
-
import
|
| 10 |
-
os.system('nvidia-smi')
|
| 11 |
print(torch.backends.cudnn.version())
|
| 12 |
|
| 13 |
def generate_filename():
|
|
@@ -126,63 +125,70 @@ def run_continuation(text, audio, model_name, chorus,
|
|
| 126 |
max_generate_audio_seconds=max_generate_audio_seconds)
|
| 127 |
return music_generation(args)
|
| 128 |
|
| 129 |
-
|
| 130 |
-
gr.
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
with gr.Row():
|
| 161 |
-
button = gr.Button("Text to Music")
|
| 162 |
-
button.click(run_text2music,
|
| 163 |
-
inputs=[text_input, model_name,
|
| 164 |
-
chorus,
|
| 165 |
-
output_sample_rate,
|
| 166 |
-
max_generate_audio_seconds],
|
| 167 |
-
outputs=music_output)
|
| 168 |
-
|
| 169 |
-
generate_button = gr.Button("Music Continuation")
|
| 170 |
-
generate_button.click(run_continuation,
|
| 171 |
-
inputs=[text_input, audio_input, model_name,
|
| 172 |
-
chorus,
|
| 173 |
-
output_sample_rate,
|
| 174 |
-
max_generate_audio_seconds],
|
| 175 |
-
outputs=music_output)
|
| 176 |
-
|
| 177 |
-
with gr.Column():
|
| 178 |
-
default_prompt_buttons = []
|
| 179 |
-
for prompt in default_prompts:
|
| 180 |
-
button = gr.Button(value=prompt)
|
| 181 |
button.click(run_text2music,
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
import datetime
|
| 7 |
import hashlib
|
| 8 |
import torch
|
| 9 |
+
from modelscope import snapshot_download
|
|
|
|
| 10 |
print(torch.backends.cudnn.version())
|
| 11 |
|
| 12 |
def generate_filename():
|
|
|
|
| 125 |
max_generate_audio_seconds=max_generate_audio_seconds)
|
| 126 |
return music_generation(args)
|
| 127 |
|
| 128 |
+
def main():
|
| 129 |
+
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
| 130 |
+
gr.Markdown("""
|
| 131 |
+
# InspireMusic
|
| 132 |
+
- Support text-to-music, music continuation, audio super-resolution, audio reconstruction tasks with high audio quality, with available sampling rates of 24kHz, 48kHz.
|
| 133 |
+
- Support long audio generation in multiple output audio formats, i.e., wav, flac, mp3, m4a.
|
| 134 |
+
- Open-source [InspireMusic-Base](https://modelscope.cn/models/iic/InspireMusic/summary), [InspireMusic-Base-24kHz](https://modelscope.cn/models/iic/InspireMusic-Base-24kHz/summary), [InspireMusic-1.5B](https://modelscope.cn/models/iic/InspireMusic-1.5B/summary), [InspireMusic-1.5B-24kHz](https://modelscope.cn/models/iic/InspireMusic-1.5B-24kHz/summary), [InspireMusic-1.5B-Long](https://modelscope.cn/models/iic/InspireMusic-1.5B-Long/summary) models for music generation.
|
| 135 |
+
- Currently only support English text prompts.
|
| 136 |
+
""")
|
| 137 |
+
|
| 138 |
+
with gr.Row(equal_height=True):
|
| 139 |
+
model_name = gr.Dropdown(["InspireMusic-1.5B-Long", "InspireMusic-1.5B", "InspireMusic-1.5B-24kHz", "InspireMusic-Base", "InspireMusic-Base-24kHz"], label="Select Model Name", value="InspireMusic-Base")
|
| 140 |
+
chorus = gr.Dropdown(["intro", "verse", "chorus", "outro"],
|
| 141 |
+
label="Chorus Mode", value="intro")
|
| 142 |
+
output_sample_rate = gr.Dropdown([48000, 24000],
|
| 143 |
+
label="Output Audio Sample Rate (Hz)",
|
| 144 |
+
value=48000)
|
| 145 |
+
max_generate_audio_seconds = gr.Slider(10, 120,
|
| 146 |
+
label="Generate Audio Length (s)",
|
| 147 |
+
value=30)
|
| 148 |
+
|
| 149 |
+
with gr.Row(equal_height=True):
|
| 150 |
+
# Textbox for custom input
|
| 151 |
+
text_input = gr.Textbox(label="Input Text (For Text-to-Music Task)", value="Experience soothing and sensual instrumental jazz with a touch of Bossa Nova, perfect for a relaxing restaurant or spa ambiance.")
|
| 152 |
+
|
| 153 |
+
audio_input = gr.Audio(label="Input Audio Prompt (For Music Continuation Task)",
|
| 154 |
+
type="filepath")
|
| 155 |
+
music_output = gr.Audio(label="Generated Music", type="filepath")
|
| 156 |
+
|
| 157 |
+
with gr.Row():
|
| 158 |
+
button = gr.Button("Text to Music")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 159 |
button.click(run_text2music,
|
| 160 |
+
inputs=[text_input, model_name,
|
| 161 |
+
chorus,
|
| 162 |
+
output_sample_rate,
|
| 163 |
+
max_generate_audio_seconds],
|
| 164 |
+
outputs=music_output)
|
| 165 |
+
|
| 166 |
+
generate_button = gr.Button("Music Continuation")
|
| 167 |
+
generate_button.click(run_continuation,
|
| 168 |
+
inputs=[text_input, audio_input, model_name,
|
| 169 |
+
chorus,
|
| 170 |
+
output_sample_rate,
|
| 171 |
+
max_generate_audio_seconds],
|
| 172 |
+
outputs=music_output)
|
| 173 |
+
|
| 174 |
+
with gr.Column():
|
| 175 |
+
default_prompt_buttons = []
|
| 176 |
+
for prompt in default_prompts:
|
| 177 |
+
button = gr.Button(value=prompt)
|
| 178 |
+
button.click(run_text2music,
|
| 179 |
+
inputs=[text_input, model_name,
|
| 180 |
+
chorus,
|
| 181 |
+
output_sample_rate,
|
| 182 |
+
max_generate_audio_seconds],
|
| 183 |
+
outputs=music_output)
|
| 184 |
+
default_prompt_buttons.append(button)
|
| 185 |
+
demo.launch()
|
| 186 |
+
|
| 187 |
+
if __name__ == '__main__':
|
| 188 |
+
model_list = ["InspireMusic-Base"]
|
| 189 |
+
for model_name in model_list:
|
| 190 |
+
if model_name == "InspireMusic-Base":
|
| 191 |
+
snapshot_download(f"iic/InspireMusic", local_dir=model_dir)
|
| 192 |
+
else:
|
| 193 |
+
snapshot_download(f"iic/{model_name}", local_dir=model_dir)os.system('nvidia-smi')
|
| 194 |
+
main()
|