Comments (2)
Environment:Tesla P40 cuda version:384.66
resnet50:
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
cifar10_128_gpu_memory | 923 | 953 | 950 | 953 | 9 | 0.947% |
cifar10_128_train_acc | 0.960 | 0.999 | 0.9837 | 0.9847 | 0.011 | 1.2% |
cifar10_128_train_speed | 548.5 | 559.5 | 553.6 | 553.5 | 3.519 | 0.635% |
flowers_64_gpu_memory | 5037 | 5541 | 5330 | 5441 | 189.98 | 3.564% |
flowers_64_train_speed | 52.59 | 53.57 | 53.23 | 53.33 | 0.286 | 0.537% |
resnet30
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
train_cost | 2.41 | 2.58 | 2.51 | 2.52 | 0.051 | 2.0% |
train_duration | 34.52 | 36.23 | 34.74 | 34.57 | 0.498 | 1.43% |
minst
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
test_acc | 0.9853 | 0.9884 | 0.9873 | 0.9874 | 0.0009 | 0.09% |
train_acc | 0.9925 | 0.9934 | 0.9930 | 0.9931 | 0.0003 | 0.03% |
train_duration | 54.05 | 54.69 | 54.36 | 54.34 | 0.022 | 0.4% |
LSTM
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
imdb_32_train_speed | 538.029 | 639.594 | 583.233 | 582.637 | 29.726 | 5.1% |
imdb_32_gpu_memory | 443 | 475 | 459 | 461 | 8.78 | 1.9% |
VGG16
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
cifar10_128_train_speed | 512.9 | 530.4 | 524.3 | 525.2 | 5.73 | 1.1% |
cifar10_128_gpu_memory | 1185 | 1185 | 1185 | 1185 | 0 | 0.0% |
flowers_32_train_speed | 26.26 | 26.76 | 26.56 | 26.60 | 0.183 | 0.68% |
flowers_32_gpu_memory | 4289 | 4443 | 4309 | 4289 | 46.96 | 1.1% |
Seq2Seq
kpi | min | max | mean | median | std | (std/mean)*100% |
---|---|---|---|---|---|---|
wmb_128_train_speed | 2685.27 | 2772.89 | 2724.99 | 2729.39 | 26.84 | 0.98% |
wmb_128_gpu_memory | 1703 | 1893 | 1804 | 1794 | 60.56 | 3.3% |
from paddle-ce-latest-kpis.
If we limit the std/mean
threshold to 2%, it seems that most factors are not stable. So can we withdraw the unstable tasks first, and add them later after some fix.
from paddle-ce-latest-kpis.
Related Issues (20)
- models repo的模型接入CE监测框架
- models repo 模型设置CE监控用的KPI阈值
- CE支持models repo模型监控方法
- model repo 待release 模型改造规范
- lstm 单卡 pass不固定 HOT 1
- debug单个模型的方法 HOT 1
- where can i import "kpi"?
- where is memory.txt
- where can i import commands HOT 1
- 所有模型去随机性 HOT 5
- seq2seq模型添加多卡
- vgg16添加多卡
- image_classification模型 acc 指标固定不下来 HOT 6
- 清理各个 task 的 log
- aws 与内网机器性能差异
- vgg16 模型random出现" Segmentation fault" HOT 1
- CE 模型负责人
- transformer模型性能下降 HOT 4
- Add distributed resnet50 model to ce
- CE部分模型使用新接口后,部分阈值需要微调
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from paddle-ce-latest-kpis.