未验证 提交 1073527c 编写于 作者: B Bin Lu 提交者: GitHub

Update feature_extraction.md

上级 2bd66c11
...@@ -85,7 +85,6 @@ Loss: ...@@ -85,7 +85,6 @@ Loss:
``` ```
python tools/train.py -c ppcls/configs/ResNet50_vd_SOP.yaml python tools/train.py -c ppcls/configs/ResNet50_vd_SOP.yaml
``` ```
- 单机多卡训练 - 单机多卡训练
``` ```
python -m paddle.distributed.launch python -m paddle.distributed.launch
...@@ -94,13 +93,11 @@ python -m paddle.distributed.launch ...@@ -94,13 +93,11 @@ python -m paddle.distributed.launch
``` ```
训练完成之后,会在`output`目录下生成`best_model` 训练完成之后,会在`output`目录下生成`best_model`
## 4.3 评估 ## 4.3 评估
- 单卡评估 - 单卡评估
``` ```
python tools/eval.py -c ppcls/configs/ResNet50_vd_SOP.yaml -o Global.pretrained_model = "output/ReModel/best_model" python tools/eval.py -c ppcls/configs/ResNet50_vd_SOP.yaml -o Global.pretrained_model = "output/ReModel/best_model"
``` ```
- 多卡评估 - 多卡评估
``` ```
python -m paddle.distributed.launch python -m paddle.distributed.launch
...@@ -108,7 +105,6 @@ python -m paddle.distributed.launch ...@@ -108,7 +105,6 @@ python -m paddle.distributed.launch
-c ppcls/configs/ResNet50_vd_SOP.yaml -c ppcls/configs/ResNet50_vd_SOP.yaml
-o Global.pretrained_model="output/ReModel/best_model" -o Global.pretrained_model="output/ReModel/best_model"
``` ```
## 4.4 推理 ## 4.4 推理
推理过程包括两个步骤: 1) 导出推理模型; 2) 获取特征向量 推理过程包括两个步骤: 1) 导出推理模型; 2) 获取特征向量
### 4.4.1 导出推理模型 ### 4.4.1 导出推理模型
......
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