Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
from TTS.api import TTS
|
2 |
import gradio as gr
|
3 |
from gradio import Dropdown
|
4 |
from scipy.io.wavfile import write
|
@@ -8,7 +8,29 @@ import re
|
|
8 |
user_choice = ""
|
9 |
MAX_NUMBER_SENTENCES = 10
|
10 |
file_upload_available = os.environ.get("ALLOW_FILE_UPLOAD")
|
11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
|
14 |
def split_process(audio, chosen_out_track):
|
@@ -30,7 +52,7 @@ def split_process(audio, chosen_out_track):
|
|
30 |
return "test.wav"
|
31 |
|
32 |
|
33 |
-
def infer(prompt, input_wav_file, clean_audio, hidden_numpy_audio):
|
34 |
print("""
|
35 |
—————
|
36 |
NEW INFERENCE:
|
@@ -79,8 +101,11 @@ NEW INFERENCE:
|
|
79 |
else:
|
80 |
prompt = prompt
|
81 |
|
|
|
|
|
|
|
82 |
gr.Info("Generating audio from prompt")
|
83 |
-
tts.tts_to_file(text=
|
84 |
file_path="output.wav",
|
85 |
voice_dir="bark_voices/",
|
86 |
speaker=f"{file_name}")
|
@@ -96,29 +121,18 @@ NEW INFERENCE:
|
|
96 |
return "output.wav", tts_video, gr.update(value=f"bark_voices/{file_name}/{contents[1]}", visible=True), gr.Group.update(visible=True), destination_path
|
97 |
|
98 |
|
99 |
-
|
100 |
-
|
101 |
-
"
|
102 |
-
"
|
103 |
-
"
|
104 |
-
|
105 |
-
|
106 |
-
positive_prompts = {
|
107 |
-
prompt_choices[0]: "I am very pleased with the progress being made to finish the cross-town transit line. This has been an excellent use of taxpayer dollars.",
|
108 |
-
prompt_choices[1]: "Yes, John is my friend. He was at my house watching the baseball game all night.",
|
109 |
-
prompt_choices[2]: "We are expecting a modest single digit increase in profits by the end of the fiscal year.",
|
110 |
-
prompt_choices[3]: "Hi Grandma it’s me, Just calling to say I love you, and I can’t wait to see you over the holidays."
|
111 |
}
|
112 |
|
113 |
-
prompt = Dropdown(
|
114 |
-
label="Text to speech prompt",
|
115 |
-
choices=prompt_choices,
|
116 |
-
elem_id="tts-prompt"
|
117 |
-
)
|
118 |
-
|
119 |
|
120 |
css = """
|
121 |
-
#col-container {max-width: 780px; margin-left: auto; margin-right: auto;}
|
|
|
122 |
a {text-decoration-line: underline; font-weight: 600;}
|
123 |
.mic-wrap > button {
|
124 |
width: 100%;
|
@@ -147,17 +161,11 @@ span.record-icon > span.dot.svelte-1thnwz {
|
|
147 |
transform: rotate(360deg);
|
148 |
}
|
149 |
}
|
150 |
-
#
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
justify-content: center;
|
156 |
-
align-items: center;
|
157 |
-
border-radius: 9999px !important;
|
158 |
-
max-width: 15rem;
|
159 |
-
height: 36px;
|
160 |
-
}
|
161 |
"""
|
162 |
|
163 |
|
@@ -166,34 +174,42 @@ def load_hidden_mic(audio_in):
|
|
166 |
return audio_in
|
167 |
|
168 |
|
169 |
-
def
|
170 |
-
|
171 |
-
|
172 |
-
|
173 |
-
|
|
|
|
|
|
|
174 |
|
175 |
|
176 |
with gr.Blocks(css=css) as demo:
|
177 |
with gr.Column(elem_id="col-container"):
|
178 |
with gr.Row():
|
179 |
with gr.Column():
|
|
|
|
|
|
|
180 |
|
181 |
-
|
182 |
-
label="
|
183 |
-
|
184 |
-
|
185 |
)
|
186 |
-
|
187 |
-
label="
|
188 |
-
|
189 |
-
|
190 |
lines=5,
|
191 |
-
|
192 |
)
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
|
|
|
|
|
197 |
|
198 |
# Replace file input with microphone input
|
199 |
micro_in = gr.Audio(
|
@@ -223,17 +239,8 @@ with gr.Blocks(css=css) as demo:
|
|
223 |
|
224 |
submit_btn.click(
|
225 |
fn=infer,
|
226 |
-
inputs=[
|
227 |
-
|
228 |
-
|
229 |
-
hidden_audio_numpy
|
230 |
-
],
|
231 |
-
outputs=[
|
232 |
-
cloned_out,
|
233 |
-
video_out,
|
234 |
-
npz_file,
|
235 |
-
folder_path
|
236 |
-
]
|
237 |
)
|
238 |
-
|
239 |
demo.queue(api_open=False, max_size=10).launch()
|
|
|
1 |
+
# from TTS.api import TTS
|
2 |
import gradio as gr
|
3 |
from gradio import Dropdown
|
4 |
from scipy.io.wavfile import write
|
|
|
8 |
user_choice = ""
|
9 |
MAX_NUMBER_SENTENCES = 10
|
10 |
file_upload_available = os.environ.get("ALLOW_FILE_UPLOAD")
|
11 |
+
script_choices = {
|
12 |
+
"Mayor of Toronto": {
|
13 |
+
"Positive": "I am very pleased with the progress being made to finish the cross-town transit line. This has been an excellent use of taxpayer dollars.",
|
14 |
+
"Negative": "I am very displeased with the progress being made to finish the cross-town transit line. This has been an embarrassing use of taxpayer dollars.",
|
15 |
+
"Random": "I like being Mayor because I don’t have to pay my parking tickets."
|
16 |
+
},
|
17 |
+
"Witness": {
|
18 |
+
"Positive": "Yes, John is my friend. He was at my house watching the baseball game all night.",
|
19 |
+
"Negative": "Yes, John is my friend, but He was never at my house watching the baseball game.",
|
20 |
+
"Random": "He is my friend, but I do not trust John."
|
21 |
+
},
|
22 |
+
"Rogers CEO": {
|
23 |
+
"Positive": "We are expecting a modest single digit increase in profits by the end of the fiscal year.",
|
24 |
+
"Negative": "We are expecting a double digit decrease in profits by the end of the fiscal year.",
|
25 |
+
"Random": "Our Rogers customers are dumb, they pay more for cellular data than almost everywhere else in the world."
|
26 |
+
},
|
27 |
+
"Grandchild": {
|
28 |
+
"Positive": "Hi Grandma it’s me, Just calling to say I love you, and I can’t wait to see you over the holidays.",
|
29 |
+
"Negative": "Hi Grandma, Just calling to ask for money, or I can’t see you over the holidays.",
|
30 |
+
"Random": "Grandma, I can’t find your email address. I need to send you something important."
|
31 |
+
}
|
32 |
+
}
|
33 |
+
# tts = TTS("tts_models/multilingual/multi-dataset/bark", gpu=True)
|
34 |
|
35 |
|
36 |
def split_process(audio, chosen_out_track):
|
|
|
52 |
return "test.wav"
|
53 |
|
54 |
|
55 |
+
def infer(prompt, input_wav_file, clean_audio, hidden_numpy_audio, script_type):
|
56 |
print("""
|
57 |
—————
|
58 |
NEW INFERENCE:
|
|
|
101 |
else:
|
102 |
prompt = prompt
|
103 |
|
104 |
+
chosen_script = script_choices.get(
|
105 |
+
theme_dropdown.get_value(), {}).get(script_type, "")
|
106 |
+
|
107 |
gr.Info("Generating audio from prompt")
|
108 |
+
tts.tts_to_file(text=chosen_script,
|
109 |
file_path="output.wav",
|
110 |
voice_dir="bark_voices/",
|
111 |
speaker=f"{file_name}")
|
|
|
121 |
return "output.wav", tts_video, gr.update(value=f"bark_voices/{file_name}/{contents[1]}", visible=True), gr.Group.update(visible=True), destination_path
|
122 |
|
123 |
|
124 |
+
# s
|
125 |
+
theme_emojis = {
|
126 |
+
"Mayor of Toronto": "🏙️",
|
127 |
+
"Witness": "👤",
|
128 |
+
"Rogers CEO": "📱",
|
129 |
+
"Grandchild": "👪"
|
|
|
|
|
|
|
|
|
|
|
|
|
130 |
}
|
131 |
|
|
|
|
|
|
|
|
|
|
|
|
|
132 |
|
133 |
css = """
|
134 |
+
#col-container {max-width: 780px; margin-left: auto; margin-right: auto; background-size: contain; background-repeat: no-repeat;}
|
135 |
+
#theme-emoji-bg {position: absolute; top: 0; left: 0; width: 100%; height: 100%; z-index: -1; opacity: 0.5; background-size: contain; background-repeat: no-repeat; background-position: center;}
|
136 |
a {text-decoration-line: underline; font-weight: 600;}
|
137 |
.mic-wrap > button {
|
138 |
width: 100%;
|
|
|
161 |
transform: rotate(360deg);
|
162 |
}
|
163 |
}
|
164 |
+
#theme-emoji {
|
165 |
+
position: absolute;
|
166 |
+
top: 10px;
|
167 |
+
right: 10px;
|
168 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
169 |
"""
|
170 |
|
171 |
|
|
|
174 |
return audio_in
|
175 |
|
176 |
|
177 |
+
def update_script_text(theme, script_type):
|
178 |
+
positive_script = script_choices.get(theme, {}).get("Positive", "")
|
179 |
+
output_script = script_choices.get(theme, {}).get(script_type, "")
|
180 |
+
|
181 |
+
# Step 3: Update the new component whenever the theme dropdown changes
|
182 |
+
theme_emoji = theme_emojis.get(theme, "")
|
183 |
+
|
184 |
+
return positive_script, output_script, theme_emoji
|
185 |
|
186 |
|
187 |
with gr.Blocks(css=css) as demo:
|
188 |
with gr.Column(elem_id="col-container"):
|
189 |
with gr.Row():
|
190 |
with gr.Column():
|
191 |
+
theme_emoji_output = gr.Label(label="Theme Emoji")
|
192 |
+
theme_dropdown = gr.Dropdown(
|
193 |
+
label="1. Select a Theme", choices=list(script_choices.keys()))
|
194 |
|
195 |
+
script_text = gr.Textbox(
|
196 |
+
label="2 & 3. Read the script below aloud THREE times for the best output:",
|
197 |
+
lines=5,
|
198 |
+
readonly=True
|
199 |
)
|
200 |
+
script_type_dropdown = gr.Dropdown(
|
201 |
+
label="4. Select the Script Type for Bot Output", choices=["Random", "Negative"])
|
202 |
+
output_script_text = gr.Textbox(
|
203 |
+
label="The bot will try to emulate the following script:",
|
204 |
lines=5,
|
205 |
+
readonly=True
|
206 |
)
|
207 |
+
theme_dropdown.change(fn=update_script_text, inputs=[
|
208 |
+
theme_dropdown, script_type_dropdown], outputs=[script_text, output_script_text])
|
209 |
+
script_type_dropdown.change(fn=update_script_text, inputs=[
|
210 |
+
theme_dropdown, script_type_dropdown], outputs=[script_text, output_script_text])
|
211 |
+
theme_dropdown.change(fn=update_script_text, inputs=[theme_dropdown, script_type_dropdown], outputs=[
|
212 |
+
script_text, output_script_text, theme_emoji_output])
|
213 |
|
214 |
# Replace file input with microphone input
|
215 |
micro_in = gr.Audio(
|
|
|
239 |
|
240 |
submit_btn.click(
|
241 |
fn=infer,
|
242 |
+
inputs=[script_text, micro_in,
|
243 |
+
hidden_audio_numpy, script_type_dropdown],
|
244 |
+
outputs=[cloned_out, video_out, npz_file, folder_path]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
245 |
)
|
|
|
246 |
demo.queue(api_open=False, max_size=10).launch()
|