Skip to content

M
models

PaddlePaddle / models
1 年多 前同步成功

通知 226
Star 6828
Fork 2962
M
models
提交 9bc02acb 编写于 作者: L LielinJiang 提交者： ruri
浏览文件
操作

Add note for ported weights of efficientNet (#3386) 

* add notes for the ported weight of EfficientNet
上级 f01ee9b1
隐藏空白更改
内联 并排
Showing with 18 addition and 14 deletion
PaddleCV/image_classification/README.md
浏览文件 @ 9bc02acb
......@@ -386,13 +386,15 @@ PaddlePaddle/Models ImageClassification 支持自定义数据
|Model | Top-1 | Top-5 | Paddle Fluid inference time(ms) | Paddle TensorRT inference time(ms) |
|- |:-: |:-: |:-: |:-: |
|[EfficientNetB0](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB0_pretrained.tar) | 77.38% | 93.31% | 10.303 | 4.334 |
|[EfficientNetB1](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB1_pretrained.tar) | 79.15% | 94.41% | 15.626 | 6.502 |
|[EfficientNetB2](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB2_pretrained.tar) | 79.85% | 94.74% | 17.847 | 7.558 |
|[EfficientNetB3](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB3_pretrained.tar) | 81.15% | 95.41% | 25.993 | 10.937 |
|[EfficientNetB4](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB4_pretrained.tar) | 82.85% | 96.23% | 47.734 | 18.536 |
|[EfficientNetB5](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB5_pretrained.tar) | 83.62% | 96.72% | 88.578 | 32.102 |
|[EfficientNetB6](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB6_pretrained.tar) | 84.00% | 96.88% | 138.670 | 51.059 |
|[EfficientNetB7](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB7_pretrained.tar) | 84.30% | 96.89% | 234.364 | 82.107 |
|[EfficientNetB1](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB1_pretrained.tar)<sup>[1](#trans)</sup> | 79.15% | 94.41% | 15.626 | 6.502 |
|[EfficientNetB2](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB2_pretrained.tar)<sup>[1](#trans)</sup> | 79.85% | 94.74% | 17.847 | 7.558 |
|[EfficientNetB3](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB3_pretrained.tar)<sup>[1](#trans)</sup> | 81.15% | 95.41% | 25.993 | 10.937 |
|[EfficientNetB4](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB4_pretrained.tar)<sup>[1](#trans)</sup> | 82.85% | 96.23% | 47.734 | 18.536 |
|[EfficientNetB5](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB5_pretrained.tar)<sup>[1](#trans)</sup> | 83.62% | 96.72% | 88.578 | 32.102 |
|[EfficientNetB6](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB6_pretrained.tar)<sup>[1](#trans)</sup> | 84.00% | 96.88% | 138.670 | 51.059 |
|[EfficientNetB7](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB7_pretrained.tar)<sup>[1](#trans)</sup> | 84.30% | 96.89% | 234.364 | 82.107 |
<a name="trans">[1]</a> 表示该预训练权重是由[官方的代码仓库](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet)转换来的。
## FAQ
......
PaddleCV/image_classification/README_en.md
浏览文件 @ 9bc02acb
......@@ -368,13 +368,15 @@ Pretrained models can be downloaded by clicking related model names.
|Model | Top-1 | Top-5 | Paddle Fluid inference time(ms) | Paddle TensorRT inference time(ms) |
|- |:-: |:-: |:-: |:-: |
|[EfficientNetB0](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB0_pretrained.tar) | 77.38% | 93.31% | 10.303 | 4.334 |
|[EfficientNetB1](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB1_pretrained.tar) | 79.15% | 94.41% | 15.626 | 6.502 |
|[EfficientNetB2](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB2_pretrained.tar) | 79.85% | 94.74% | 17.847 | 7.558 |
|[EfficientNetB3](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB3_pretrained.tar) | 81.15% | 95.41% | 25.993 | 10.937 |
|[EfficientNetB4](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB4_pretrained.tar) | 82.85% | 96.23% | 47.734 | 18.536 |
|[EfficientNetB5](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB5_pretrained.tar) | 83.62% | 96.72% | 88.578 | 32.102 |
|[EfficientNetB6](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB6_pretrained.tar) | 84.00% | 96.88% | 138.670 | 51.059 |
|[EfficientNetB7](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB7_pretrained.tar) | 84.30% | 96.89% | 234.364 | 82.107 |
|[EfficientNetB1](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB1_pretrained.tar)<sup>[1](#trans)</sup> | 79.15% | 94.41% | 15.626 | 6.502 |
|[EfficientNetB2](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB2_pretrained.tar)<sup>[1](#trans)</sup> | 79.85% | 94.74% | 17.847 | 7.558 |
|[EfficientNetB3](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB3_pretrained.tar)<sup>[1](#trans)</sup> | 81.15% | 95.41% | 25.993 | 10.937 |
|[EfficientNetB4](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB4_pretrained.tar)<sup>[1](#trans)</sup> | 82.85% | 96.23% | 47.734 | 18.536 |
|[EfficientNetB5](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB5_pretrained.tar)<sup>[1](#trans)</sup> | 83.62% | 96.72% | 88.578 | 32.102 |
|[EfficientNetB6](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB6_pretrained.tar)<sup>[1](#trans)</sup> | 84.00% | 96.88% | 138.670 | 51.059 |
|[EfficientNetB7](https://paddle-imagenet-models-name.bj.bcebos.com/EfficientNetB7_pretrained.tar)<sup>[1](#trans)</sup> | 84.30% | 96.89% | 234.364 | 82.107 |
<a name="trans">[1]</a> means the pretrained weight is converted form [original repository](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet).
## FAQ
......
Markdown is supported
0% 或 .
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 或