Spaces:
Paused
Paused
show Llama prompt + share button
Browse files
app.py
CHANGED
|
@@ -9,6 +9,7 @@ from gradio_client import Client
|
|
| 9 |
client = Client("https://fffiloni-test-llama-api.hf.space/", hf_token=hf_token)
|
| 10 |
lyrics_client = Client("https://fffiloni-music-to-lyrics.hf.space/")
|
| 11 |
|
|
|
|
| 12 |
|
| 13 |
from compel import Compel, ReturnedEmbeddingsType
|
| 14 |
from diffusers import DiffusionPipeline
|
|
@@ -92,7 +93,7 @@ def infer(audio_file, has_lyrics):
|
|
| 92 |
|
| 93 |
Here's the music description :
|
| 94 |
|
| 95 |
-
Β«
|
| 96 |
|
| 97 |
And here are the lyrics :
|
| 98 |
|
|
@@ -109,7 +110,7 @@ def infer(audio_file, has_lyrics):
|
|
| 109 |
|
| 110 |
Here's the music description :
|
| 111 |
|
| 112 |
-
Β«
|
| 113 |
"""
|
| 114 |
print("""βββ
|
| 115 |
Calling Llama2 ...
|
|
@@ -123,7 +124,7 @@ def infer(audio_file, has_lyrics):
|
|
| 123 |
|
| 124 |
print(f"Llama2 result: {result}")
|
| 125 |
|
| 126 |
-
#
|
| 127 |
print("""βββ
|
| 128 |
Calling SD-XL ...
|
| 129 |
""")
|
|
@@ -134,10 +135,83 @@ def infer(audio_file, has_lyrics):
|
|
| 134 |
print("Finished")
|
| 135 |
|
| 136 |
#return cap_result, result, images
|
| 137 |
-
return images, result, gr.update(visible=True)
|
| 138 |
|
| 139 |
css = """
|
| 140 |
-
#col-container {max-width:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 141 |
"""
|
| 142 |
with gr.Blocks(css=css) as demo:
|
| 143 |
with gr.Column(elem_id="col-container"):
|
|
@@ -161,23 +235,51 @@ with gr.Blocks(css=css) as demo:
|
|
| 161 |
Note: Only the first 30 seconds of your audio will be used for inference.
|
| 162 |
</p>
|
| 163 |
</div>""")
|
|
|
|
| 164 |
audio_input = gr.Audio(label="Music input", type="filepath", source="upload")
|
| 165 |
-
|
|
|
|
|
|
|
| 166 |
infer_btn = gr.Button("Generate Image from Music")
|
| 167 |
#lpmc_cap = gr.Textbox(label="Lp Music Caps caption")
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 171 |
|
| 172 |
gr.Examples(examples=[["./examples/electronic.mp3", "No"],["./examples/folk.wav", "No"], ["./examples/orchestra.wav", "No"]],
|
| 173 |
fn=infer,
|
| 174 |
inputs=[audio_input, has_lyrics],
|
| 175 |
-
outputs=[img_result, llama_trans_cap, tryagain_btn],
|
| 176 |
cache_examples=True
|
| 177 |
)
|
| 178 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 179 |
#infer_btn.click(fn=infer, inputs=[audio_input], outputs=[lpmc_cap, llama_trans_cap, img_result])
|
| 180 |
-
infer_btn.click(fn=infer, inputs=[audio_input, has_lyrics], outputs=[img_result, llama_trans_cap, tryagain_btn])
|
|
|
|
| 181 |
tryagain_btn.click(fn=solo_xd, inputs=[llama_trans_cap], outputs=[img_result])
|
| 182 |
|
| 183 |
demo.queue(max_size=20).launch()
|
|
|
|
| 9 |
client = Client("https://fffiloni-test-llama-api.hf.space/", hf_token=hf_token)
|
| 10 |
lyrics_client = Client("https://fffiloni-music-to-lyrics.hf.space/")
|
| 11 |
|
| 12 |
+
from share_btn import community_icon_html, loading_icon_html, share_js
|
| 13 |
|
| 14 |
from compel import Compel, ReturnedEmbeddingsType
|
| 15 |
from diffusers import DiffusionPipeline
|
|
|
|
| 93 |
|
| 94 |
Here's the music description :
|
| 95 |
|
| 96 |
+
Β« {cap_result} Β»
|
| 97 |
|
| 98 |
And here are the lyrics :
|
| 99 |
|
|
|
|
| 110 |
|
| 111 |
Here's the music description :
|
| 112 |
|
| 113 |
+
Β« {cap_result} Β»
|
| 114 |
"""
|
| 115 |
print("""βββ
|
| 116 |
Calling Llama2 ...
|
|
|
|
| 124 |
|
| 125 |
print(f"Llama2 result: {result}")
|
| 126 |
|
| 127 |
+
# βββ
|
| 128 |
print("""βββ
|
| 129 |
Calling SD-XL ...
|
| 130 |
""")
|
|
|
|
| 135 |
print("Finished")
|
| 136 |
|
| 137 |
#return cap_result, result, images
|
| 138 |
+
return images, result, gr.update(visible=True), gr.Group.update(visible=True)
|
| 139 |
|
| 140 |
css = """
|
| 141 |
+
#col-container {max-width: 780px; margin-left: auto; margin-right: auto;}
|
| 142 |
+
a {text-decoration-line: underline; font-weight: 600;}
|
| 143 |
+
.animate-spin {
|
| 144 |
+
animation: spin 1s linear infinite;
|
| 145 |
+
}
|
| 146 |
+
@keyframes spin {
|
| 147 |
+
from {
|
| 148 |
+
transform: rotate(0deg);
|
| 149 |
+
}
|
| 150 |
+
to {
|
| 151 |
+
transform: rotate(360deg);
|
| 152 |
+
}
|
| 153 |
+
}
|
| 154 |
+
#share-btn-container {
|
| 155 |
+
display: flex;
|
| 156 |
+
padding-left: 0.5rem !important;
|
| 157 |
+
padding-right: 0.5rem !important;
|
| 158 |
+
background-color: #000000;
|
| 159 |
+
justify-content: center;
|
| 160 |
+
align-items: center;
|
| 161 |
+
border-radius: 9999px !important;
|
| 162 |
+
max-width: 13rem;
|
| 163 |
+
}
|
| 164 |
+
div#share-btn-container > div {
|
| 165 |
+
flex-direction: row;
|
| 166 |
+
background: black;
|
| 167 |
+
align-items: center;
|
| 168 |
+
}
|
| 169 |
+
#share-btn-container:hover {
|
| 170 |
+
background-color: #060606;
|
| 171 |
+
}
|
| 172 |
+
#share-btn {
|
| 173 |
+
all: initial;
|
| 174 |
+
color: #ffffff;
|
| 175 |
+
font-weight: 600;
|
| 176 |
+
cursor:pointer;
|
| 177 |
+
font-family: 'IBM Plex Sans', sans-serif;
|
| 178 |
+
margin-left: 0.5rem !important;
|
| 179 |
+
padding-top: 0.5rem !important;
|
| 180 |
+
padding-bottom: 0.5rem !important;
|
| 181 |
+
right:0;
|
| 182 |
+
}
|
| 183 |
+
#share-btn * {
|
| 184 |
+
all: unset;
|
| 185 |
+
}
|
| 186 |
+
#share-btn-container div:nth-child(-n+2){
|
| 187 |
+
width: auto !important;
|
| 188 |
+
min-height: 0px !important;
|
| 189 |
+
}
|
| 190 |
+
#share-btn-container .wrap {
|
| 191 |
+
display: none !important;
|
| 192 |
+
}
|
| 193 |
+
#share-btn-container.hidden {
|
| 194 |
+
display: none!important;
|
| 195 |
+
}
|
| 196 |
+
.footer {
|
| 197 |
+
margin-bottom: 45px;
|
| 198 |
+
margin-top: 10px;
|
| 199 |
+
text-align: center;
|
| 200 |
+
border-bottom: 1px solid #e5e5e5;
|
| 201 |
+
}
|
| 202 |
+
.footer>p {
|
| 203 |
+
font-size: .8rem;
|
| 204 |
+
display: inline-block;
|
| 205 |
+
padding: 0 10px;
|
| 206 |
+
transform: translateY(10px);
|
| 207 |
+
background: white;
|
| 208 |
+
}
|
| 209 |
+
.dark .footer {
|
| 210 |
+
border-color: #303030;
|
| 211 |
+
}
|
| 212 |
+
.dark .footer>p {
|
| 213 |
+
background: #0b0f19;
|
| 214 |
+
}
|
| 215 |
"""
|
| 216 |
with gr.Blocks(css=css) as demo:
|
| 217 |
with gr.Column(elem_id="col-container"):
|
|
|
|
| 235 |
Note: Only the first 30 seconds of your audio will be used for inference.
|
| 236 |
</p>
|
| 237 |
</div>""")
|
| 238 |
+
|
| 239 |
audio_input = gr.Audio(label="Music input", type="filepath", source="upload")
|
| 240 |
+
with gr.Row():
|
| 241 |
+
has_lyrics = gr.Radio(label="Does your audio has lyrics ?", choices=["Yes", "No"], value="No", info="If yes, the image should reflect the lyrics, but be aware that because we add a step (getting lyrics), inference will take more time.")
|
| 242 |
+
song_title = gr.Textbox(label="Song Title", value="Title: ", interactive=True, info="If you want to share your result, please provide the title of your audio sample :)", elem_id="song-title")
|
| 243 |
infer_btn = gr.Button("Generate Image from Music")
|
| 244 |
#lpmc_cap = gr.Textbox(label="Lp Music Caps caption")
|
| 245 |
+
with gr.Row():
|
| 246 |
+
llama_trans_cap = gr.Textbox(label="Llama Image Suggestion", placeholder="Llama2 image prompt suggestion will be displayed here ;)", visible=True, lines=12, elem_id="llama-prompt")
|
| 247 |
+
img_result = gr.Image(label="Image Result", elem_id="image-out")
|
| 248 |
+
with gr.Row():
|
| 249 |
+
tryagain_btn = gr.Button("Try another image ?", visible=False)
|
| 250 |
+
with gr.Group(elem_id="share-btn-container", visible=False) as share_group:
|
| 251 |
+
community_icon = gr.HTML(community_icon_html)
|
| 252 |
+
loading_icon = gr.HTML(loading_icon_html)
|
| 253 |
+
share_button = gr.Button("Share to community", elem_id="share-btn")
|
| 254 |
|
| 255 |
gr.Examples(examples=[["./examples/electronic.mp3", "No"],["./examples/folk.wav", "No"], ["./examples/orchestra.wav", "No"]],
|
| 256 |
fn=infer,
|
| 257 |
inputs=[audio_input, has_lyrics],
|
| 258 |
+
outputs=[img_result, llama_trans_cap, tryagain_btn, share_group],
|
| 259 |
cache_examples=True
|
| 260 |
)
|
| 261 |
|
| 262 |
+
gr.HTML("""
|
| 263 |
+
<div class="footer">
|
| 264 |
+
<p>
|
| 265 |
+
Music to Image Demo by π€ <a href="https://twitter.com/fffiloni" target="_blank">Sylvain Filoni</a>
|
| 266 |
+
</p>
|
| 267 |
+
</div>
|
| 268 |
+
<div id="may-like-container" style="display: flex;justify-content: center;flex-direction: column;align-items: center;">
|
| 269 |
+
<p style="font-size: 0.8em;margin-bottom: 4px;">You may also like: </p>
|
| 270 |
+
<div id="may-like" style="display:flex; align-items:center; justify-content: center;height:20px;">
|
| 271 |
+
<svg height="20" width="182" style="margin-left:4px">
|
| 272 |
+
<a href="https://huggingface.co/spaces/fffiloni/Music-To-Zeroscope" target="_blank">
|
| 273 |
+
<image href="https://img.shields.io/badge/π€ Spaces-Music To Zeroscope-blue" src="https://img.shields.io/badge/π€ Spaces-Music To Zeroscope-blue.png" height="20"/>
|
| 274 |
+
</a>
|
| 275 |
+
</svg>
|
| 276 |
+
</div>
|
| 277 |
+
</div>
|
| 278 |
+
""")
|
| 279 |
+
|
| 280 |
#infer_btn.click(fn=infer, inputs=[audio_input], outputs=[lpmc_cap, llama_trans_cap, img_result])
|
| 281 |
+
infer_btn.click(fn=infer, inputs=[audio_input, has_lyrics], outputs=[img_result, llama_trans_cap, tryagain_btn, share_group])
|
| 282 |
+
share_button.click(None, [], [], _js=share_js)
|
| 283 |
tryagain_btn.click(fn=solo_xd, inputs=[llama_trans_cap], outputs=[img_result])
|
| 284 |
|
| 285 |
demo.queue(max_size=20).launch()
|