Spaces:
Paused
Paused
File size: 7,656 Bytes
296c11e eb21a2f 58ffcfd bb9d46e 296c11e eb21a2f bb9d46e eb21a2f bb9d46e 296c11e bb9d46e eb21a2f 296c11e bb9d46e eb21a2f bb9d46e c9cdb67 296c11e c9cdb67 eb21a2f c9cdb67 296c11e c9cdb67 296c11e c9cdb67 296c11e c9cdb67 eb21a2f 296c11e c9cdb67 296c11e c9cdb67 eb21a2f 296c11e eb21a2f c9cdb67 296c11e eb21a2f 296c11e 58ffcfd 296c11e 58ffcfd c9cdb67 58ffcfd c9cdb67 296c11e eb21a2f c9cdb67 eb21a2f c9cdb67 eb21a2f bb9d46e c9cdb67 7af48a1 bb9d46e 296c11e c9cdb67 bb9d46e eb21a2f bb9d46e 58ffcfd bb9d46e 58ffcfd bb9d46e 58ffcfd bb9d46e c9cdb67 296c11e c9cdb67 bb9d46e 296c11e bb9d46e 296c11e bb9d46e c9cdb67 bb9d46e 296c11e 58ffcfd |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 |
import json
import gradio as gr
from gradio import Dropdown
# from share_btn import community_icon_html, loading_icon_html, share_js
import os
import shutil
import re
user_choice = ""
# with open("characters.json", "r") as file:
# data = json.load(file)
# characters = [
# {
# "image": item["image"],
# "title": item["title"],
# "speaker": item["speaker"]
# }
# for item in data
# ]
# tts = TTS("tts_models/multilingual/multi-dataset/bark", gpu=True)
# def update_selection(selected_state: gr.SelectData):
# c_image = characters[selected_state.index]["image"]
# c_title = characters[selected_state.index]["title"]
# c_speaker = characters[selected_state.index]["speaker"]
# return c_title, selected_state
def infer(prompt, input_wav_file, clean_audio, hidden_numpy_audio):
print("""
—————
NEW INFERENCE:
———————
""")
if prompt == "":
gr.Warning("Do not forget to provide a tts prompt !")
if clean_audio is True:
print("We want to clean audio sample")
new_name = os.path.splitext(os.path.basename(input_wav_file))[0]
if os.path.exists(os.path.join("bark_voices", f"{new_name}_cleaned")):
print("This file has already been cleaned")
check_name = os.path.join("bark_voices", f"{new_name}_cleaned")
source_path = os.path.join(check_name, f"{new_name}_cleaned.wav")
else:
source_path = split_process(hidden_numpy_audio, "vocals")
new_path = os.path.join(os.path.dirname(
source_path), f"{new_name}_cleaned.wav")
os.rename(source_path, new_path)
source_path = new_path
else:
source_path = input_wav_file
destination_directory = "bark_voices"
file_name = os.path.splitext(os.path.basename(source_path))[0]
destination_path = os.path.join(destination_directory, file_name)
os.makedirs(destination_path, exist_ok=True)
shutil.move(source_path, os.path.join(
destination_path, f"{file_name}.wav"))
sentences = re.split(r'(?<=[.!?])\s+', prompt)
if len(sentences) > MAX_NUMBER_SENTENCES:
gr.Info("Your text is too long. To keep this demo enjoyable for everyone, we only kept the first 10 sentences :) Duplicate this space and set MAX_NUMBER_SENTENCES for longer texts ;)")
first_nb_sentences = sentences[:MAX_NUMBER_SENTENCES]
limited_prompt = ' '.join(first_nb_sentences)
prompt = limited_prompt
else:
prompt = prompt
gr.Info("Generating audio from prompt")
tts.tts_to_file(text=prompt,
file_path="output.wav",
voice_dir="bark_voices/",
speaker=f"{file_name}")
contents = os.listdir(f"bark_voices/{file_name}")
for item in contents:
print(item)
print("Preparing final waveform video ...")
tts_video = gr.make_waveform(audio="output.wav")
print(tts_video)
print("FINISHED")
return "output.wav", tts_video, gr.update(value=f"bark_voices/{file_name}/{contents[1]}", visible=True), gr.Group.update(visible=True), destination_path
prompt_choices = [
"I am very displeased with the progress being made to finish the cross-town transit line. transit line. This has been an embarrassing use of taxpayer dollars.",
"Yes, John is my friend, but He was never at my house watching the baseball game.",
"We are expecting a double digit increase in profits by the end of the fiscal year.",
"Hi Grandma, Just calling to ask for money, or I can't see you over the holidays. "
]
positive_prompts = {
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.",
prompt_choices[1]: "Yes, John is my friend. He was at my house watching the baseball game all night.",
prompt_choices[2]: "We are expecting a modest single digit increase in profits by the end of the fiscal year.",
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."
}
prompt = Dropdown(
label="Text to speech prompt",
choices=prompt_choices,
elem_id="tts-prompt"
)
css = """
#col-container {max-width: 780px; margin-left: auto; margin-right: auto;}
a {text-decoration-line: underline; font-weight: 600;}
.mic-wrap > button {
width: 100%;
height: 60px;
font-size: 1.4em!important;
}
.record-icon.svelte-1thnwz {
display: flex;
position: relative;
margin-right: var(--size-2);
width: unset;
height: unset;
}
span.record-icon > span.dot.svelte-1thnwz {
width: 20px!important;
height: 20px!important;
}
.animate-spin {
animation: spin 1s linear infinite;
}
@keyframes spin {
from {
transform: rotate(0deg);
}
to {
transform: rotate(360deg);
}
}
#share-btn-container {
display: flex;
padding-left: 0.5rem !important;
padding-right: 0.5rem !important;
background-color: #000000;
justify-content: center;
align-items: center;
border-radius: 9999px !important;
max-width: 15rem;
height: 36px;
}
"""
def load_hidden_mic(audio_in):
print("USER RECORDED A NEW SAMPLE")
return audio_in
def update_positive_prompt(prompt_value):
global user_choice
user_choice = prompt_value
if prompt_value in positive_prompts:
return positive_prompts[prompt_value]
with gr.Blocks(css=css) as demo:
with gr.Column(elem_id="col-container"):
with gr.Row():
with gr.Column():
prompt = gr.Dropdown(
label="Negative Speech Prompt",
choices=prompt_choices,
elem_id="tts-prompt"
)
texts_samples = gr.Textbox(
label="Positive prompts",
info="Please read out this prompt 5 times to generate a good sample",
value="",
lines=5,
elem_id="texts_samples"
)
# Connect the prompt change to the update_positive_prompt function
prompt.change(fn=update_positive_prompt,
inputs=prompt, outputs=texts_samples)
# Replace file input with microphone input
micro_in = gr.Audio(
label="Record voice to clone",
type="filepath",
source="microphone",
interactive=True
)
hidden_audio_numpy = gr.Audio(type="numpy", visible=False)
submit_btn = gr.Button("Submit")
with gr.Column():
cloned_out = gr.Audio(
label="Text to speech output", visible=False)
video_out = gr.Video(label="Waveform video",
elem_id="voice-video-out")
npz_file = gr.File(label=".npz file", visible=False)
folder_path = gr.Textbox(visible=False)
micro_in.stop_recording(fn=load_hidden_mic, inputs=[micro_in], outputs=[
hidden_audio_numpy], queue=False)
submit_btn.click(
fn=infer,
inputs=[
prompt,
micro_in,
hidden_audio_numpy
],
outputs=[
cloned_out,
video_out,
npz_file,
folder_path
]
)
demo.queue(api_open=False, max_size=10).launch()
|