Spaces:
Running
Running
File size: 5,077 Bytes
0d80816 ea32d9a 0d80816 b07c30e 0d80816 ea32d9a b37a421 ea32d9a b07c30e 0d80816 b07c30e 0d80816 ea32d9a 0d80816 8129b0f b07c30e 0d80816 90c437b 0d80816 ea32d9a 0d80816 b07c30e 90c437b b07c30e 0d80816 b07c30e 0d80816 b07c30e 0d80816 9da2dd2 0d80816 2afaead b07c30e 90c437b b07c30e 90c437b b07c30e 90c437b 0d80816 b07c30e 0d80816 b07c30e 1768f9f b07c30e 90c437b b07c30e 90c437b b07c30e 90c437b b07c30e a91dcdc b07c30e a91dcdc 825451d b07c30e 0d80816 b07c30e 0d80816 b07c30e |
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 |
# Copyright (c) 2023 Amphion.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree
import subprocess
command_to_run = "cd ./modules/monotonic_align;mkdir -p monotonic_align;python setup.py build_ext --inplace;cd /home/user/app"
subprocess.check_output(command_to_run, shell=True, text=True)
import gradio as gr
import os
import inference
SUPPORTED_SPEAKERS = {
"Cori Samuel":"hifitts_92",
"Phil Benson":"hifitts_6097",
"Mike Pelton":"hifitts_6670",
"Tony Oliva":"hifitts_6671",
"Maria Kasper":"hifitts_8051",
"John Van Stan":"hifitts_9017",
"Helen Taylor":"hifitts_9136",
"Sylviamb":"hifitts_11614",
"Celine Major":"hifitts_11697",
"LikeManyWaters":"hifitts_12787"
}
def tts_inference(
input_text,
target_speaker,
duration
):
### Target Speaker ###
target_speaker = SUPPORTED_SPEAKERS[target_speaker]
args_list = ["--config", "./egs/tts/vits_hifitts/exp_config.json"]
args_list += ["--checkpoint_path", "./expdir/checkpoint/latest-checkpoint"]
args_list += ["--speaker_name_1", target_speaker]
args_list += ["--speaker_name_2", None]
args_list += ["--text", input_text]
args_list += ["--mode","single"]
args_list += ["--duration_control",str(float(2.05-duration))]
args_list += ["--output_dir", "result"]
args_list += ["--log_level", "debug"]
os.environ["WORK_DIR"] = "./"
inference.main(args_list)
### Display ###
result_file = os.path.join(
"result/single/test_pred.wav"
)
return result_file
def tc_inference(
input_text,
target_speaker_1,
target_speaker_2,
confusion_degree,
duration
):
### Target Speaker ###
target_speaker_1 = SUPPORTED_SPEAKERS[target_speaker_1]
if target_speaker_2 is not None:
target_speaker_2 = SUPPORTED_SPEAKERS[target_speaker_2]
args_list = ["--config", "./egs/tts/vits_hifitts/exp_config.json"]
args_list += ["--checkpoint_path", "./expdir/checkpoint/latest-checkpoint"]
args_list += ["--speaker_name_1", target_speaker_1]
args_list += ["--speaker_name_2", target_speaker_2]
args_list += ["--alpha", str(float(confusion_degree))]
args_list += ["--text", input_text]
args_list += ["--mode","single"]
args_list += ["--duration_control",str(float(2.05-duration))]
args_list += ["--output_dir", "result"]
args_list += ["--log_level", "debug"]
os.environ["WORK_DIR"] = "./"
inference.main(args_list)
### Display ###
source_speaker_1 = os.path.join(
"result/single/s1.wav"
)
source_speaker_2 = os.path.join(
"result/single/s2.wav"
)
result_file = os.path.join(
"result/single/test_pred.wav"
)
return source_speaker_1, source_speaker_2, result_file
# Section 1: TTS
tts_demo_inputs = [
gr.Textbox(
label="Input Text",
type="text",
placeholder="Type something here.."
),
gr.Radio(
choices=list(SUPPORTED_SPEAKERS.keys()),
label="Target Speaker",
value="Cori Samuel"
),
gr.Slider(
0.1,
2,
value=1,
step=0.05,
label="Speaking Rate",
info="As the step number increases, the speaking rate will be faster.",
)
]
tts_demo_output = gr.Audio(label="Generated Speech")
# Section 2: Timbre confusion
tc_demo_inputs = [
gr.Textbox(
label="Input Text",
type="text",
placeholder="Type something here.."
),
gr.Radio(
choices=list(SUPPORTED_SPEAKERS.keys()),
label="Target Speaker 1",
value="Cori Samuel"
),
gr.Radio(
choices=list(SUPPORTED_SPEAKERS.keys()),
label="Target Speaker 2",
value="Phil Benson"
),
gr.Slider(
0,
1,
value=0.5,
step=0.1,
label="Fusion Degree",
info="As the step number increases, the generated voice will be more similar to speaker 2.",
),
gr.Slider(
0.1,
2,
value=1,
step=0.05,
label="Speaking Rate",
info="As the step number increases, the speaking rate will be faster.",
)
]
tc_demo_outputs = [
gr.Audio(label="Target Speaker 1"),
gr.Audio(label="Target Speaker 2"),
gr.Audio(label="Interpolated Speech")
]
with gr.Blocks() as demo:
gr.Interface(
fn=tts_inference,
inputs=tts_demo_inputs,
outputs=tts_demo_output,
title="Amphion Text-to-Speech",
description="This demo offers an Amphion TTS pretrained model (VITS) for you to explore."
)
gr.Interface(
fn=tc_inference,
inputs=tc_demo_inputs,
outputs=tc_demo_outputs,
title="Voice Fusion",
description="In this section, you can choose two speakers to create a voice mix. Adjust the ‘Fusion Degree’ slider to customize your desired mix ratio between the two speakers."
)
demo.queue()
demo.launch()
# if __name__ == "__main__":
# demo.launch(share=True) |