diff --git a/docs/zh_CN_tmp/image_recognition_pipeline/feature_extraction.md b/docs/zh_CN_tmp/image_recognition_pipeline/feature_extraction.md index 17db708fe705418217884aac1c4f16be1c0f145b..17b980f20bbd57ed2ba26f876c788339e9e02cd6 100644 --- a/docs/zh_CN_tmp/image_recognition_pipeline/feature_extraction.md +++ b/docs/zh_CN_tmp/image_recognition_pipeline/feature_extraction.md @@ -113,14 +113,12 @@ python -m paddle.distributed.launch 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 ``` -python tools/export_model -c xxx -o Global.pretrained_model = xxxx +python tools/export_model -c ppcls/configs/ResNet50_vd_SOP.yaml -o Global.pretrained_model = "output/ReModel/best_model" ``` 生成的推理模型位于inference目录,名字为inference.pd* ### 4.4.2 获取特征向量 ``` cd deploy -python python/inference_rec.py -c configs/ O rec_inference_model_dir: "../inference/inference" -``` -输出 +python python/inference_rec.py -c configs/inference_product.yaml -o rec_inference_model_dir="../inference/inference" ```