Spaces:
Runtime error
Runtime error
File size: 5,158 Bytes
890de26 10d71a1 890de26 2d21960 890de26 5e7226b 5cbdda6 10d71a1 890de26 5e7226b 890de26 5cbdda6 10d71a1 e32152a 890de26 |
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 |
import os
import gradio as gr
import numpy as np
import soundfile as sf
import torchaudio
from speechbrain.pretrained.interfaces import foreign_class
from app_utils import preprocess_video_and_rank
from authors import AUTHORS
# Importing necessary components for the Gradio app
from description import DESCRIPTION_DYNAMIC # , DESCRIPTION_STATIC
# import scipy.io.wavfile as wav
from paraformer import AudioReader, CttPunctuator, FSMNVad, ParaformerOffline
os.environ["no_proxy"] = "localhost,127.0.0.1,::1"
###########################语音部分######################################
classifier = foreign_class(
source="pretrained_models/local-speechbrain/emotion-recognition-wav2vec2-IEMOCAP", # ".\\emotion-recognition-wav2vec2-IEMOCAP"
pymodule_file="custom_interface.py",
classname="CustomEncoderWav2vec2Classifier",
savedir="pretrained_models/local-speechbrain/emotion-recognition-wav2vec2-IEMOCAP",
)
ASR_model = ParaformerOffline()
vad = FSMNVad()
punc = CttPunctuator()
def classify_continuous(audio):
print(type(audio))
print(audio)
sample_rate, signal = audio # 这是语音的输入
signal = signal.astype(np.float32)
signal /= np.max(np.abs(signal))
sf.write("a.wav", signal, sample_rate)
signal, sample_rate = torchaudio.load("a.wav")
signal1 = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)(
signal
)
torchaudio.save("out.wav", signal1, 16000, encoding="PCM_S", bits_per_sample=16)
Audio = "out.wav"
speech, sample_rate = AudioReader.read_wav_file(Audio)
if signal == "none":
return "none", "none", "haha"
else:
segments = vad.segments_offline(speech)
text_results = ""
for part in segments:
_result = ASR_model.infer_offline(
speech[part[0] * 16 : part[1] * 16], hot_words="任意热词 空格分开"
)
text_results += punc.punctuate(_result)[0]
out_prob, score, index, text_lab = classifier.classify_batch(signal1)
return text_results, out_prob.squeeze(0).numpy(), text_lab[-1], Audio
#########################################视频部分###################################
def clear_dynamic_info():
return (
gr.Video(value=None),
gr.Plot(value=None),
gr.Textbox(""),
)
##################################设置各自的app类####################
with gr.Blocks(css="app.css") as video:
with gr.Tab("Dynamic App"):
gr.Markdown(value=DESCRIPTION_DYNAMIC)
with gr.Row():
with gr.Column(scale=2):
input_video = gr.Video(
sources=["webcam", "upload"], elem_classes="video1", format='mp4'
)
with gr.Row():
clear_btn_dynamic = gr.Button(
value="Clear", interactive=True, scale=1
)
# submit_dynamic = gr.Button(
# value="Submit", interactive=True, scale=1, elem_classes="submit"
# )
submit_and_rank = gr.Button(
value="Score", interactive=True, scale=1, elem_classes="submit"
)
with gr.Column(scale=2, elem_classes="dl4"):
with gr.Row():
output_score = gr.Textbox(label="scores")
output_statistics = gr.Plot(
label="Statistics of emotions", elem_classes="stat"
)
output_audio=gr.Audio(interactive=False)
audio_test_button=gr.Button("分析语音")
out1=gr.Textbox(label="语音分析结果")
out2=gr.Textbox(label="音频情感识别1")
out3=gr.Textbox(label="音频情感识别2")
internal_audio=gr.Audio(interactive=False)
gr.Examples(
[
"videos/video1.mp4",
"videos/video2.mp4",
"videos/sample.webm",
"videos/cnm.mp4",
],
[input_video],
)
with gr.Tab("Authors"):
gr.Markdown(value=AUTHORS)
clear_btn_dynamic.click(
fn=clear_dynamic_info,
inputs=[],
outputs=[
input_video,
output_statistics,
output_score,
],
queue=True,
)
submit_and_rank.click(
fn=preprocess_video_and_rank,
inputs=input_video,
outputs=[
output_statistics,
output_score,
output_audio,
],
)
audio_test_button.click(
fn=classify_continuous,
inputs=output_audio,
outputs=[out1,out2,out3,internal_audio]
)
####################################
speech = gr.Interface(
classify_continuous,
gr.Audio(sources=["microphone"]),
[
gr.Text(label="语音识别结果"),
gr.Text(label="音频情感识别1"),
gr.Text(label="音频情感识别2"),
],
)
with gr.Blocks() as app:
with gr.Tab("语音"):
speech.render()
with gr.Tab("视频"):
video.render()
app.launch()
|