Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -10,13 +10,13 @@ from transformers import (
|
|
10 |
MusicgenForConditionalGeneration
|
11 |
)
|
12 |
from io import BytesIO
|
13 |
-
from streamlit_lottie import st_lottie
|
14 |
|
15 |
# ---------------------------------------------------------------------
|
16 |
# 1) PAGE CONFIG
|
17 |
# ---------------------------------------------------------------------
|
18 |
st.set_page_config(
|
19 |
-
page_title="Radio Imaging AI
|
20 |
page_icon="π§",
|
21 |
layout="wide"
|
22 |
)
|
@@ -26,26 +26,24 @@ st.set_page_config(
|
|
26 |
# ---------------------------------------------------------------------
|
27 |
CUSTOM_CSS = """
|
28 |
<style>
|
29 |
-
/*
|
30 |
body {
|
31 |
background-color: #121212;
|
32 |
color: #FFFFFF;
|
33 |
font-family: "Helvetica Neue", sans-serif;
|
34 |
}
|
35 |
|
36 |
-
/* Main container width */
|
37 |
.block-container {
|
38 |
max-width: 1100px;
|
39 |
padding: 1rem 1.5rem;
|
40 |
}
|
41 |
|
42 |
-
/* Headings with a neon-ish green accent */
|
43 |
h1, h2, h3 {
|
44 |
-
color: #1DB954;
|
45 |
margin-bottom: 0.5rem;
|
46 |
}
|
47 |
|
48 |
-
/*
|
49 |
.stButton>button {
|
50 |
background-color: #1DB954 !important;
|
51 |
color: #FFFFFF !important;
|
@@ -59,13 +57,12 @@ h1, h2, h3 {
|
|
59 |
background-color: #1ed760 !important;
|
60 |
}
|
61 |
|
62 |
-
/* Sidebar: black background
|
63 |
.sidebar .sidebar-content {
|
64 |
background-color: #000000;
|
65 |
color: #FFFFFF;
|
66 |
}
|
67 |
|
68 |
-
/* Text inputs and text areas */
|
69 |
textarea, input, select {
|
70 |
border-radius: 8px !important;
|
71 |
background-color: #282828 !important;
|
@@ -73,20 +70,20 @@ textarea, input, select {
|
|
73 |
border: 1px solid #3e3e3e;
|
74 |
}
|
75 |
|
76 |
-
/* Audio
|
77 |
audio {
|
78 |
width: 100%;
|
79 |
margin-top: 1rem;
|
80 |
}
|
81 |
|
82 |
-
/* Lottie container
|
83 |
.lottie-container {
|
84 |
display: flex;
|
85 |
justify-content: center;
|
86 |
margin-bottom: 20px;
|
87 |
}
|
88 |
|
89 |
-
/* Footer
|
90 |
.footer-note {
|
91 |
text-align: center;
|
92 |
font-size: 14px;
|
@@ -94,31 +91,83 @@ audio {
|
|
94 |
margin-top: 2rem;
|
95 |
}
|
96 |
|
97 |
-
/* Hide Streamlit
|
98 |
#MainMenu, footer {visibility: hidden;}
|
99 |
</style>
|
100 |
"""
|
101 |
st.markdown(CUSTOM_CSS, unsafe_allow_html=True)
|
102 |
|
103 |
# ---------------------------------------------------------------------
|
104 |
-
# 3)
|
105 |
# ---------------------------------------------------------------------
|
106 |
@st.cache_data
|
107 |
def load_lottie_url(url: str):
|
108 |
-
"""
|
109 |
-
Fetch Lottie JSON for animations.
|
110 |
-
"""
|
111 |
r = requests.get(url)
|
112 |
if r.status_code != 200:
|
113 |
return None
|
114 |
return r.json()
|
115 |
|
116 |
-
# Example Lottie animation (radio waves / music eq, etc.)
|
117 |
LOTTIE_URL = "https://assets3.lottiefiles.com/temp/lf20_Q6h5zV.json"
|
118 |
lottie_animation = load_lottie_url(LOTTIE_URL)
|
119 |
|
120 |
# ---------------------------------------------------------------------
|
121 |
-
# 4)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
122 |
# ---------------------------------------------------------------------
|
123 |
with st.sidebar:
|
124 |
st.header("π Radio Library")
|
@@ -131,21 +180,19 @@ with st.sidebar:
|
|
131 |
st.markdown("<br>", unsafe_allow_html=True)
|
132 |
|
133 |
# ---------------------------------------------------------------------
|
134 |
-
#
|
135 |
# ---------------------------------------------------------------------
|
136 |
col1, col2 = st.columns([3, 2], gap="large")
|
137 |
|
138 |
with col1:
|
139 |
-
st.title("AI Radio Imaging
|
140 |
-
st.subheader("
|
141 |
|
142 |
st.markdown(
|
143 |
"""
|
144 |
-
Create **radio imaging promos** and **jingles** with
|
145 |
-
|
146 |
-
|
147 |
-
- Uses Meta's [MusicGen](https://github.com/facebookresearch/audiocraft) for **audio**.
|
148 |
-
- Features a Spotify-like UI & Lottie animations for a modern user experience.
|
149 |
"""
|
150 |
)
|
151 |
with col2:
|
@@ -158,152 +205,86 @@ with col2:
|
|
158 |
st.markdown("---")
|
159 |
|
160 |
# ---------------------------------------------------------------------
|
161 |
-
#
|
162 |
# ---------------------------------------------------------------------
|
163 |
-
st.subheader("π Step 1:
|
164 |
|
165 |
prompt = st.text_area(
|
166 |
-
"
|
167 |
height=120
|
168 |
)
|
169 |
|
170 |
col_model, col_device = st.columns(2)
|
171 |
with col_model:
|
172 |
llama_model_id = st.text_input(
|
173 |
-
"Llama Model
|
174 |
-
value="meta-llama/Llama-3
|
175 |
-
help="
|
176 |
)
|
177 |
with col_device:
|
178 |
device_option = st.selectbox(
|
179 |
-
"
|
180 |
["auto", "cpu"],
|
181 |
-
help="
|
182 |
)
|
183 |
|
184 |
-
# ---------------------------------------------------------------------
|
185 |
-
# 7) BUTTON: GENERATE RADIO SCRIPT WITH LLAMA
|
186 |
-
# ---------------------------------------------------------------------
|
187 |
if st.button("π Generate Promo Script"):
|
188 |
if not prompt.strip():
|
189 |
-
st.error("Please
|
190 |
else:
|
191 |
-
with st.spinner("Generating script..."):
|
192 |
try:
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
refined_text = generate_radio_script(prompt, pipeline_llama)
|
197 |
-
st.session_state["refined_script"] = refined_text
|
198 |
st.success("Promo script generated!")
|
199 |
-
st.write(
|
200 |
except Exception as e:
|
201 |
-
st.error(f"
|
202 |
|
203 |
st.markdown("---")
|
204 |
|
205 |
# ---------------------------------------------------------------------
|
206 |
-
#
|
207 |
# ---------------------------------------------------------------------
|
208 |
-
st.subheader("πΆ Step 2: Generate
|
209 |
|
210 |
-
|
211 |
|
212 |
if st.button("π§ Create Audio with MusicGen"):
|
213 |
-
|
214 |
-
|
215 |
-
st.error("Please generate a promo script first.")
|
216 |
else:
|
217 |
-
with st.spinner("
|
218 |
try:
|
219 |
-
# Load MusicGen
|
220 |
mg_model, mg_processor = load_musicgen_model()
|
221 |
-
|
222 |
-
|
223 |
-
# Prepare model input
|
224 |
inputs = mg_processor(
|
225 |
-
text=[
|
226 |
-
|
227 |
-
|
228 |
)
|
229 |
-
|
230 |
-
audio_values = mg_model.generate(**inputs, max_new_tokens=audio_tokens)
|
231 |
sr = mg_model.config.audio_encoder.sampling_rate
|
232 |
-
|
233 |
-
# Save audio to WAV
|
234 |
-
out_filename = "radio_imaging_output.wav"
|
235 |
-
scipy.io.wavfile.write(out_filename, rate=sr, data=audio_values[0,0].numpy())
|
236 |
-
|
237 |
-
st.success("Audio created! Press play to listen:")
|
238 |
-
st.audio(out_filename)
|
239 |
-
except Exception as e:
|
240 |
-
st.error(f"Error generating audio: {e}")
|
241 |
-
|
242 |
-
# ---------------------------------------------------------------------
|
243 |
-
# 9) HELPER FUNCTIONS
|
244 |
-
# ---------------------------------------------------------------------
|
245 |
-
@st.cache_resource
|
246 |
-
def load_llama_pipeline(model_id: str, device: str):
|
247 |
-
"""
|
248 |
-
Load the Llama model & pipeline.
|
249 |
-
"""
|
250 |
-
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
251 |
-
model = AutoModelForCausalLM.from_pretrained(
|
252 |
-
model_id,
|
253 |
-
torch_dtype=torch.float16 if device == "auto" else torch.float32,
|
254 |
-
device_map=device
|
255 |
-
)
|
256 |
-
text_gen_pipeline = pipeline(
|
257 |
-
"text-generation",
|
258 |
-
model=model,
|
259 |
-
tokenizer=tokenizer,
|
260 |
-
device_map=device
|
261 |
-
)
|
262 |
-
return text_gen_pipeline
|
263 |
|
264 |
-
|
265 |
-
|
266 |
-
Use Llama to refine the user's input into a brief but creative radio imaging script.
|
267 |
-
"""
|
268 |
-
system_prompt = (
|
269 |
-
"You are a top-tier radio imaging producer. "
|
270 |
-
"Take the user's concept and craft a short, high-impact promo script. "
|
271 |
-
"Include style, tone, and potential CTA if relevant."
|
272 |
-
)
|
273 |
-
full_prompt = f"{system_prompt}\nUser concept: {user_input}\nRefined script:"
|
274 |
-
|
275 |
-
output = pipeline_llama(
|
276 |
-
full_prompt,
|
277 |
-
max_new_tokens=200,
|
278 |
-
do_sample=True,
|
279 |
-
temperature=0.9
|
280 |
-
)[0]["generated_text"]
|
281 |
-
|
282 |
-
# Attempt to isolate the final script portion
|
283 |
-
if "Refined script:" in output:
|
284 |
-
output = output.split("Refined script:", 1)[-1].strip()
|
285 |
-
output += "\n\n(Generated by Llama in Radio Imaging MVP)"
|
286 |
-
|
287 |
-
return output
|
288 |
|
289 |
-
|
290 |
-
|
291 |
-
|
292 |
-
|
293 |
-
"""
|
294 |
-
mg_model = MusicgenForConditionalGeneration.from_pretrained("facebook/musicgen-small")
|
295 |
-
mg_processor = AutoProcessor.from_pretrained("facebook/musicgen-small")
|
296 |
-
return mg_model, mg_processor
|
297 |
|
298 |
# ---------------------------------------------------------------------
|
299 |
-
#
|
300 |
# ---------------------------------------------------------------------
|
301 |
st.markdown("---")
|
302 |
st.markdown(
|
303 |
"""
|
304 |
<div class="footer-note">
|
305 |
-
|
306 |
-
|
307 |
</div>
|
308 |
""",
|
309 |
unsafe_allow_html=True
|
|
|
10 |
MusicgenForConditionalGeneration
|
11 |
)
|
12 |
from io import BytesIO
|
13 |
+
from streamlit_lottie import st_lottie
|
14 |
|
15 |
# ---------------------------------------------------------------------
|
16 |
# 1) PAGE CONFIG
|
17 |
# ---------------------------------------------------------------------
|
18 |
st.set_page_config(
|
19 |
+
page_title="Radio Imaging AI with Llama 3",
|
20 |
page_icon="π§",
|
21 |
layout="wide"
|
22 |
)
|
|
|
26 |
# ---------------------------------------------------------------------
|
27 |
CUSTOM_CSS = """
|
28 |
<style>
|
29 |
+
/* Dark background with Spotify-like vibe */
|
30 |
body {
|
31 |
background-color: #121212;
|
32 |
color: #FFFFFF;
|
33 |
font-family: "Helvetica Neue", sans-serif;
|
34 |
}
|
35 |
|
|
|
36 |
.block-container {
|
37 |
max-width: 1100px;
|
38 |
padding: 1rem 1.5rem;
|
39 |
}
|
40 |
|
|
|
41 |
h1, h2, h3 {
|
42 |
+
color: #1DB954;
|
43 |
margin-bottom: 0.5rem;
|
44 |
}
|
45 |
|
46 |
+
/* Rounded, bright green button on hover */
|
47 |
.stButton>button {
|
48 |
background-color: #1DB954 !important;
|
49 |
color: #FFFFFF !important;
|
|
|
57 |
background-color: #1ed760 !important;
|
58 |
}
|
59 |
|
60 |
+
/* Sidebar: black background */
|
61 |
.sidebar .sidebar-content {
|
62 |
background-color: #000000;
|
63 |
color: #FFFFFF;
|
64 |
}
|
65 |
|
|
|
66 |
textarea, input, select {
|
67 |
border-radius: 8px !important;
|
68 |
background-color: #282828 !important;
|
|
|
70 |
border: 1px solid #3e3e3e;
|
71 |
}
|
72 |
|
73 |
+
/* Audio styling */
|
74 |
audio {
|
75 |
width: 100%;
|
76 |
margin-top: 1rem;
|
77 |
}
|
78 |
|
79 |
+
/* Lottie container */
|
80 |
.lottie-container {
|
81 |
display: flex;
|
82 |
justify-content: center;
|
83 |
margin-bottom: 20px;
|
84 |
}
|
85 |
|
86 |
+
/* Footer */
|
87 |
.footer-note {
|
88 |
text-align: center;
|
89 |
font-size: 14px;
|
|
|
91 |
margin-top: 2rem;
|
92 |
}
|
93 |
|
94 |
+
/* Hide Streamlit branding if you wish */
|
95 |
#MainMenu, footer {visibility: hidden;}
|
96 |
</style>
|
97 |
"""
|
98 |
st.markdown(CUSTOM_CSS, unsafe_allow_html=True)
|
99 |
|
100 |
# ---------------------------------------------------------------------
|
101 |
+
# 3) LOAD LOTTIE ANIMATION
|
102 |
# ---------------------------------------------------------------------
|
103 |
@st.cache_data
|
104 |
def load_lottie_url(url: str):
|
|
|
|
|
|
|
105 |
r = requests.get(url)
|
106 |
if r.status_code != 200:
|
107 |
return None
|
108 |
return r.json()
|
109 |
|
|
|
110 |
LOTTIE_URL = "https://assets3.lottiefiles.com/temp/lf20_Q6h5zV.json"
|
111 |
lottie_animation = load_lottie_url(LOTTIE_URL)
|
112 |
|
113 |
# ---------------------------------------------------------------------
|
114 |
+
# 4) LOAD LLAMA 3 (GATED MODEL) - WITH use_auth_token
|
115 |
+
# ---------------------------------------------------------------------
|
116 |
+
@st.cache_resource
|
117 |
+
def load_llama_pipeline(model_id: str, device: str):
|
118 |
+
"""
|
119 |
+
Load the Llama 3 model from Hugging Face.
|
120 |
+
Requires huggingface-cli login if model is gated.
|
121 |
+
"""
|
122 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id, use_auth_token=True)
|
123 |
+
model = AutoModelForCausalLM.from_pretrained(
|
124 |
+
model_id,
|
125 |
+
torch_dtype=torch.float16 if device == "auto" else torch.float32,
|
126 |
+
device_map=device,
|
127 |
+
use_auth_token=True
|
128 |
+
)
|
129 |
+
text_gen_pipeline = pipeline(
|
130 |
+
"text-generation",
|
131 |
+
model=model,
|
132 |
+
tokenizer=tokenizer,
|
133 |
+
device_map=device
|
134 |
+
)
|
135 |
+
return text_gen_pipeline
|
136 |
+
|
137 |
+
# ---------------------------------------------------------------------
|
138 |
+
# 5) REFINE SCRIPT (LLAMA)
|
139 |
+
# ---------------------------------------------------------------------
|
140 |
+
def generate_radio_script(user_input: str, pipeline_llama) -> str:
|
141 |
+
system_prompt = (
|
142 |
+
"You are a top-tier radio imaging producer using Llama 3. "
|
143 |
+
"Take the user's concept and craft a short, creative promo script."
|
144 |
+
)
|
145 |
+
combined_prompt = f"{system_prompt}\nUser concept: {user_input}\nRefined script:"
|
146 |
+
|
147 |
+
result = pipeline_llama(
|
148 |
+
combined_prompt,
|
149 |
+
max_new_tokens=200,
|
150 |
+
do_sample=True,
|
151 |
+
temperature=0.9
|
152 |
+
)
|
153 |
+
output_text = result[0]["generated_text"]
|
154 |
+
|
155 |
+
if "Refined script:" in output_text:
|
156 |
+
output_text = output_text.split("Refined script:", 1)[-1].strip()
|
157 |
+
output_text += "\n\n(Generated by Llama 3 - Radio Imaging)"
|
158 |
+
return output_text
|
159 |
+
|
160 |
+
# ---------------------------------------------------------------------
|
161 |
+
# 6) LOAD MUSICGEN
|
162 |
+
# ---------------------------------------------------------------------
|
163 |
+
@st.cache_resource
|
164 |
+
def load_musicgen_model():
|
165 |
+
mg_model = MusicgenForConditionalGeneration.from_pretrained("facebook/musicgen-small")
|
166 |
+
mg_processor = AutoProcessor.from_pretrained("facebook/musicgen-small")
|
167 |
+
return mg_model, mg_processor
|
168 |
+
|
169 |
+
# ---------------------------------------------------------------------
|
170 |
+
# 7) SIDEBAR
|
171 |
# ---------------------------------------------------------------------
|
172 |
with st.sidebar:
|
173 |
st.header("π Radio Library")
|
|
|
180 |
st.markdown("<br>", unsafe_allow_html=True)
|
181 |
|
182 |
# ---------------------------------------------------------------------
|
183 |
+
# 8) HEADER
|
184 |
# ---------------------------------------------------------------------
|
185 |
col1, col2 = st.columns([3, 2], gap="large")
|
186 |
|
187 |
with col1:
|
188 |
+
st.title("AI Radio Imaging with Llama 3")
|
189 |
+
st.subheader("Gated Model + MusicGen Audio")
|
190 |
|
191 |
st.markdown(
|
192 |
"""
|
193 |
+
Create **radio imaging promos** and **jingles** with Llama 3 + MusicGen.
|
194 |
+
**Note**: You must have access to `"meta-llama/Llama-3-70B-Instruct"` on Hugging Face,
|
195 |
+
and be logged in via `huggingface-cli login`.
|
|
|
|
|
196 |
"""
|
197 |
)
|
198 |
with col2:
|
|
|
205 |
st.markdown("---")
|
206 |
|
207 |
# ---------------------------------------------------------------------
|
208 |
+
# 9) SCRIPT GENERATION
|
209 |
# ---------------------------------------------------------------------
|
210 |
+
st.subheader("π Step 1: Describe Your Promo Idea")
|
211 |
|
212 |
prompt = st.text_area(
|
213 |
+
"Example: 'A 15-second hype jingle for a morning talk show, fun and energetic.'",
|
214 |
height=120
|
215 |
)
|
216 |
|
217 |
col_model, col_device = st.columns(2)
|
218 |
with col_model:
|
219 |
llama_model_id = st.text_input(
|
220 |
+
"Llama 3 Model ID",
|
221 |
+
value="meta-llama/Llama-3-70B-Instruct", # Official ID if you have it
|
222 |
+
help="Use the exact name you see on the Hugging Face model page."
|
223 |
)
|
224 |
with col_device:
|
225 |
device_option = st.selectbox(
|
226 |
+
"Device (GPU vs CPU)",
|
227 |
["auto", "cpu"],
|
228 |
+
help="If you have GPU, 'auto' tries to use it; CPU might be slow."
|
229 |
)
|
230 |
|
|
|
|
|
|
|
231 |
if st.button("π Generate Promo Script"):
|
232 |
if not prompt.strip():
|
233 |
+
st.error("Please type some concept first.")
|
234 |
else:
|
235 |
+
with st.spinner("Generating script with Llama 3..."):
|
236 |
try:
|
237 |
+
llm_pipeline = load_llama_pipeline(llama_model_id, device_option)
|
238 |
+
final_script = generate_radio_script(prompt, llm_pipeline)
|
239 |
+
st.session_state["final_script"] = final_script
|
|
|
|
|
240 |
st.success("Promo script generated!")
|
241 |
+
st.write(final_script)
|
242 |
except Exception as e:
|
243 |
+
st.error(f"Llama generation error: {e}")
|
244 |
|
245 |
st.markdown("---")
|
246 |
|
247 |
# ---------------------------------------------------------------------
|
248 |
+
# 10) AUDIO GENERATION: MUSICGEN
|
249 |
# ---------------------------------------------------------------------
|
250 |
+
st.subheader("πΆ Step 2: Generate Audio")
|
251 |
|
252 |
+
audio_length = st.slider("MusicGen Max Tokens (approx track length)", 128, 1024, 512, 64)
|
253 |
|
254 |
if st.button("π§ Create Audio with MusicGen"):
|
255 |
+
if "final_script" not in st.session_state:
|
256 |
+
st.error("No script found. Please generate a script first.")
|
|
|
257 |
else:
|
258 |
+
with st.spinner("Creating audio..."):
|
259 |
try:
|
|
|
260 |
mg_model, mg_processor = load_musicgen_model()
|
261 |
+
text_for_audio = st.session_state["final_script"]
|
262 |
+
|
|
|
263 |
inputs = mg_processor(
|
264 |
+
text=[text_for_audio],
|
265 |
+
padding=True,
|
266 |
+
return_tensors="pt"
|
267 |
)
|
268 |
+
audio_values = mg_model.generate(**inputs, max_new_tokens=audio_length)
|
|
|
269 |
sr = mg_model.config.audio_encoder.sampling_rate
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
270 |
|
271 |
+
outfile = "llama3_radio_jingle.wav"
|
272 |
+
scipy.io.wavfile.write(outfile, rate=sr, data=audio_values[0, 0].numpy())
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
273 |
|
274 |
+
st.success("Audio generated! Press play below:")
|
275 |
+
st.audio(outfile)
|
276 |
+
except Exception as e:
|
277 |
+
st.error(f"MusicGen error: {e}")
|
|
|
|
|
|
|
|
|
278 |
|
279 |
# ---------------------------------------------------------------------
|
280 |
+
# 11) FOOTER
|
281 |
# ---------------------------------------------------------------------
|
282 |
st.markdown("---")
|
283 |
st.markdown(
|
284 |
"""
|
285 |
<div class="footer-note">
|
286 |
+
Β© 2025 Radio Imaging with Llama 3 β Built using Hugging Face & Streamlit. <br>
|
287 |
+
Log in via <code>huggingface-cli</code> and ensure access to <strong>meta-llama/Llama-3-70B-Instruct</strong>.
|
288 |
</div>
|
289 |
""",
|
290 |
unsafe_allow_html=True
|