Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
ModelZoo
aasist_bladedisc
Commits
8a32a0df
Commit
8a32a0df
authored
Nov 27, 2023
by
zhangxiao-stack
Browse files
修改格式
parent
f65189bc
Changes
3
Hide whitespace changes
Inline
Side-by-side
Showing
3 changed files
with
40 additions
and
26 deletions
+40
-26
Aassist_Backbone.PNG
Aassist_Backbone.PNG
+0
-0
README.md
README.md
+35
-23
model.properties
model.properties
+5
-3
No files found.
Aassist_Backbone.PNG
0 → 100644
View file @
8a32a0df
284 KB
README.md
View file @
8a32a0df
...
...
@@ -3,29 +3,18 @@
开源的音频反欺骗的模型
## 模型结构
aasist
,参考链接https://github.com/clovaai/aasist, 参考论文ASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks
aasist
是一种开源的音频反欺诈的模型,主要的模型结构如下所示:
## 数据集
脚本下载方式:
```
python ./download_dataset.py
```
手动下载方式:
```
ASVspoof2019 dataset: https://datashare.ed.ac.uk/handle/10283/3336
```
下载LA.zip文件,unzip解压
## 推理

### 环境配置
提供
[
光源
](
https://www.sourcefind.cn/#/service-details
)
拉取的训练的docker镜像:
## 环境配置
### Docker(方法一)
提供
[
光源
](
https://www.sourcefind.cn/#/service-list
)
拉取的训练的docker镜像:
*
推理镜像:
```
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:aasist-main
```
激活镜像环境:
*
激活镜像环境:
```
source /root/env_disc.sh
cd /root/aasist;sh run.sh
...
...
@@ -34,7 +23,20 @@ cd /root/aasist;sh run.sh
```
pip3 install -r requirement.txt
```
### 推理命令,使用BladeDISC进行加速
## 数据集
脚本下载方式:
```
python ./download_dataset.py
```
手动下载方式:
```
ASVspoof2019 dataset: https://datashare.ed.ac.uk/handle/10283/3336
```
下载LA.zip文件,unzip解压
## 推理
To evaluate AASIST [1]:
```
...
...
@@ -51,14 +53,24 @@ export TORCH_MHLO_OP_WHITE_LIST="aten::max;aten::batch_norm;aten::abs,aten::selu
python3 main.py --eval --config ./config/AASIST-L.conf
python3 main_opt.py --eval --config ./config/AASIST-L.conf
```
### 单卡测试
测试命令:
```
bash run.sh
```
## 精度
使用Blade DISC优化后的精度与未使用Blade DISC优化后的精度保持一致
## 应用场景
### 算法类别
语音识别
### 热点行业
金融,交通,教育
### 源码仓库及问题反馈
https://developer.hpccube.com/codes/modelzoo/bladedisc_aasist
##
源码仓库及问题反馈
*
https://
developer.hpcc
ub
e
.com/c
odes/modelzoo/bladedisc_
aasist
##
# 参考
https://
gith
ub.com/c
lovaai/
aasist
.git
## 参考
*
https://github.com/clovaai/aasist.git
model.properties
View file @
8a32a0df
# 模型唯一标识
modelCode
=
135
# 模型名称
modelName
=
AASIST_BladeDISC
modelName
=
aasist_bladedisc
# 模型描述
modelDescription
=
AASIST是开源的音频反欺骗的模型,基于BladeDISC进行深度学习编译优化
# 应用场景
(多个标签以英文逗号分割)
# 应用场景
appScenario
=
推理,NLP,语音模型
# 框架类型(多个标签以英文逗号分割)
frameType
=
BladeDISC,Py
t
orch
frameType
=
BladeDISC,Py
T
orch
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment