# Copyright (c) Microsoft
# 2022 Chengdong Liang (liangchengdong@mail.nwpu.edu.cn)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import gradio as gr
import wespeaker
STYLE = """
"""
OUTPUT_OK = (STYLE + """
The speakers are
{:.1f}%
similar
Welcome, human!
(You must get at least 70% to be considered the same person)
""")
OUTPUT_FAIL = (STYLE + """
The speakers are
{:.1f}%
similar
Warning! stranger!
(You must get at least 70% to be considered the same person)
""")
OUTPUT_ERROR = (STYLE + """
Input Error
{}!
""")
cn_model = wespeaker.load_model("chinese")
en_model = wespeaker.load_model("english")
def speaker_verification(audio_path1, audio_path2, lang='CN'):
if audio_path1 == None or audio_path2 == None:
output = OUTPUT_ERROR.format('Please enter two audios')
return output
if lang == 'EN':
model = en_model
elif lang == 'CN':
model = cn_model
else:
output = OUTPUT_ERROR.format('Please select a language')
return output
cos_score = model.compute_similarity(audio_path1, audio_path2)
if cos_score >= 0.70:
output = OUTPUT_OK.format(cos_score * 100)
else:
output = OUTPUT_FAIL.format(cos_score * 100)
return output
# input
inputs = [
gr.inputs.Audio(source="microphone",
type="filepath",
optional=True,
label='Speaker#1'),
gr.inputs.Audio(source="microphone",
type="filepath",
optional=True,
label='Speaker#2'),
gr.Radio(['EN', 'CN'], label='Language'),
]
output = gr.outputs.HTML(label="")
# description
description = ("
WeSpeaker Demo ! Try it with your own voice ! Note: We recommend that the audio length be greater than 5s !