提交 7161e13e 编写于 作者: I itminner 提交者: whs

add mobilenetv1 resnet50 quantization results (#3456)

上级 e9c7c30e
......@@ -125,12 +125,22 @@ mobile预测模型兼容Paddle-Lite(Paddle-Mobile的升级版), 使用方法
| weight量化方式 | activation量化方式| top1_acc/top5_acc |Paddle Fluid inference time(ms)| Paddle Lite inference time(ms)|
|---|---|---|---|---|
|baseline|- |70.99%/89.68%|- |-|
|abs_max|abs_max|- |- |-|
|abs_max|moving_average_abs_max|- |- |-|
|channel_wise_abs_max|abs_max|- |- |-|
|abs_max|abs_max|70.74%/89.55% |- |-|
|abs_max|moving_average_abs_max|70.89%/89.67% |- |-|
|channel_wise_abs_max|abs_max|70.93%/89.65% |- |-|
>训练超参:
优化器
```
fluid.optimizer.Momentum(momentum=0.9,
learning_rate=fluid.layers.piecewise_decay(
boundaries=[5000 * 12],
values=[0.0001, 0.00001]),
regularization=fluid.regularizer.L2Decay(1e-4))
```
batch size 1024
### MobileNetV2
| weight量化方式 | activation量化方式| top1_acc/top5_acc |Paddle Fluid inference time(ms)| Paddle Lite inference time(ms)|
......@@ -147,10 +157,20 @@ mobile预测模型兼容Paddle-Lite(Paddle-Mobile的升级版), 使用方法
| weight量化方式 | activation量化方式| top1_acc/top5_acc |Paddle Fluid inference time(ms)| Paddle Lite inference time(ms)|
|---|---|---|---|---|
|baseline|- |76.50%/93.00%|- |-|
|abs_max|abs_max|- |- |-|
|abs_max|moving_average_abs_max|- |- |-|
|channel_wise_abs_max|abs_max|- |- |-|
|abs_max|abs_max|76.71%/93.10% |- |-|
|abs_max|moving_average_abs_max|76.54%/93.12% |- |-|
|channel_wise_abs_max|abs_max|76.56%/93.05% |- |-|
>训练超参:
优化器
```
fluid.optimizer.Momentum(momentum=0.9,
learning_rate=fluid.layers.piecewise_decay(
boundaries=[5000 * 12],
values=[0.0001, 0.00001]),
regularization=fluid.regularizer.L2Decay(1e-4))
```
batch size 1024
## FAQ
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