提交 978396e7 编写于 作者: T typhoonzero

Merge branch 'dist_train_benchmark_vgg16' of...

Merge branch 'dist_train_benchmark_vgg16' of https://github.com/typhoonzero/Paddle into dist_train_benchmark_vgg16
...@@ -48,16 +48,18 @@ ...@@ -48,16 +48,18 @@
### Different Pserver Count ### Different Pserver Count
- Trainer Count: 100 - Trainer Count: 60
- Batch Size: 128 - Batch Size: 128
- Metrics: mini-batch / sec - Metrics: mini-batch / sec
| PServer Count | 10 | 20 | 40 | 60 | | PServer Count | 3 | 6 |10 | 20 |
| -- | -- | -- | -- | -- | | -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | - | - | - | - | | PaddlePaddle Fluid(should fix in next PR) | 589.1 | 592.6 | 656.4 | 655.8 |
| PaddlePaddle v2 | - | - | - | - | | PaddlePaddle v2 | 593.4 | 791.3 | 729.7 | 821.7 |
| TensorFlow | - | - | - | - | | TensorFlow | - | - | - | - |
*The performance gap between Fuild and v2 comes from the network interference.*
## Steps to run the performance test ## Steps to run the performance test
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册