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
MLPerf_BERT_paddle
Commits
9294c4ab
Commit
9294c4ab
authored
Apr 26, 2023
by
liangjing
Browse files
add model.properties
parent
53a3fabf
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
14 additions
and
6 deletions
+14
-6
README.md
README.md
+5
-6
model.properties
model.properties
+9
-0
No files found.
README.md
View file @
9294c4ab
...
...
@@ -34,6 +34,10 @@ BERT-large包含24个Transformer编码器,每个编码器有1024个隐藏层
## 训练
### 测试规模
单机8卡进行性能&&精度测试
### 环境配置
提供
[
光源
](
https://www.sourcefind.cn/#/service-details
)
拉取的训练的docker镜像:
*
训练镜像:
...
...
@@ -42,7 +46,7 @@ python依赖安装:
pip3 install -r requirement.txt
### 训练
训练命令
(此处以单机8卡规模为例说明)
:
训练命令:
bash run_8gpu.sh
...
...
@@ -54,11 +58,6 @@ python依赖安装:
├── /workpalce/bert_data/phase1
└── └──model.ckpt-28252.tf_pickled #预训练模型
### 单卡测试
测试命令:
bash run_1gpu.sh
## 性能和准确率数据
测试采用上述输入数据,加速卡采用Z100L,可最终达到收敛要求;
...
...
model.properties
0 → 100644
View file @
9294c4ab
# 模型名称
modelName
=
Bert-large
# 模型描述
modelDescription
=
Bert-large是一种基于Transformer架构的预训练语言模型
# 应用场景(多个标签以英文逗号分割)
appScenario
=
NLP,训练
# 框架类型(多个标签以英文逗号分割)
frameType
=
Paddle
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