Closed lguowang closed 2 years ago
heavengate
commented
2 years ago {'err_no': 8, 'err_msg': "(data_id=0 log_id=0) [ppyolo_v3|0] Failed to postprocess: 'scale_factor.lod'", 'key': [], 'value': [], 'tensors': []}
看你报错用的还是ppyolo的部署配置吧，可以参考文档重新导出一下picodet的模型 https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.3/deploy/serving#3-导出模型
lguowang
commented
2 years ago {'err_no': 8, 'err_msg': "(data_id=0 log_id=0) [ppyolo_v3|0] Failed to postprocess: 'scale_factor.lod'", 'key': [], 'value': [], 'tensors': []}
看你报错用的还是ppyolo的部署配置吧，可以参考文档重新导出一下picodet的模型 https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.3/deploy/serving#3-导出模型
我们项目需要用 web_service 进行部署，有相关的参考链接吗？ 需要picodet的，不是ppyolo的
你好，我用pp2.3训练了PicoDet，然后pp(develop)导出模型进行sering部署， 但预测时出现主要的错误：{'err_no': 8, 'err_msg': "(data_id=0 log_id=0) [ppyolo_v3|0] Failed to postprocess: 'scale_factor.lod'", 'key': [], 'value': [], 'tensors': []}
请问我要怎么修改才能让PicoDet在Sering上运行？
说明：服务和客户端参考的是Serving/examples/Pipeline/PaddleDetection/ppyolo_mbv3/下的 web_service.py 和pipeline_http_client.py
serving_server 中的 serving_server_conf.prototxt ：
feed_var { name: "image" alias_name: "image" is_lod_tensor: false feed_type: 1 shape: 3 shape: 416 shape: 416 } fetch_var { name: "save_infer_model/scale_0.tmp_1" alias_name: "save_infer_model/scale_0.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_1.tmp_1" alias_name: "save_infer_model/scale_1.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_2.tmp_1" alias_name: "save_infer_model/scale_2.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_3.tmp_1" alias_name: "save_infer_model/scale_3.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_4.tmp_1" alias_name: "save_infer_model/scale_4.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_5.tmp_1" alias_name: "save_infer_model/scale_5.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_6.tmp_1" alias_name: "save_infer_model/scale_6.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 } fetch_var { name: "save_infer_model/scale_7.tmp_1" alias_name: "save_infer_model/scale_7.tmp_1" is_lod_tensor: true fetch_type: 1 shape: -1 }
我配置的config.yml:
dag: is_thread_op: false tracer: interval_s: 20 http_port: 18888 op: yolov3: concurrency: 10 local_service_conf: client_type: local_predictor device_type: 1 devices: '2' fetch_list:
请问我要怎么修改才能让PicoDet在Sering上运行？