Serving icon indicating copy to clipboard operation
Serving copied to clipboard

A flexible, high-performance carrier for machine learning models(『飞桨』服务化部署框架)

Results 218 Serving issues
Sort by recently updated
recently updated
newest added

pip列表: paddle-bfloat 0.1.7 paddle-serving-app 0.3.0 paddle-serving-client 0.5.0 paddle-serving-server 0.5.0 paddle-serving-server-gpu 0.5.0.post11 paddlepaddle-gpu 2.3.2.post116 压力测试一段时间之后,内存溢出

怎样关闭模型预测时的服务日志 PipelineClient::predict pack_data time:1680858497.491558 环境: paddle-serving-app 0.9.0 paddle-serving-client 0.9.0 paddle-serving-server-gpu 0.8.3.post102

使用[链接](https://github.com/PaddlePaddle/Serving/blob/v0.9.0/examples/Pipeline/PaddleDetection/yolov3/README_CN.md)中官方给的模型,无法跑通yolov3:先启动web_service.py(启动正常),然后启动pipeline_http_client.py,将报错如下: ``` Traceback (most recent call last): File "/usr/local/lib/python3.6/site-packages/paddle_serving_server/pipeline/error_catch.py", line 97, in wrapper res = func(*args, **kw) File "/usr/local/lib/python3.6/site-packages/paddle_serving_server/pipeline/operator.py", line 1181, in postprocess_help logid_dict.get(data_id)) File "web_service.py", line 64, in postprocess...

Bumps [json](https://github.com/douglascrockford/JSON-java) from 20190722 to 20230227. Release notes Sourced from json's releases. 20230227 Pull Request Description #723 Protect JSONML from stack overflow exceptions caused by recursion #720 Limit the XML...

dependencies
java

使用paddleserver部署paddleCls,如何输出top5? 使用python predict_cls.py可以输出top5,但使用paddlerserver部署后只能输出top1,部署命令是python classification_web_service.py 。 在哪里可以修改输出类别数??

环境: PaddleVideo版本:V2.2.2 paddle-serving-app:0.9.0 paddle-serving-client : 0.9.0 paddle-serving-server-gpu: 0.9.0.post1028 paddlepaddle-gpu : 2.4.2.post116 项目: 当前使用FootballAction进行模型训练,训练lstm使用train_lstm,导出使用train_lstm下的export_inference_model.py。最终,模型导出视图如下: ![58e3b151cc2076962b606412309a1c8](https://user-images.githubusercontent.com/35001756/231431816-fb90c2df-3104-45a5-a450-f35aadbd89a8.png) 使用paddle_serving_client.convert去将这个模型转换成paddle_serving_server要求的模型。模型信息如下: ![1a09895bddcac9918baeb3bbf56e93a](https://user-images.githubusercontent.com/35001756/231431916-69461878-bd21-4937-9c88-eeba6e9ee314.png) 在部署Pipeline后,预测报出的错误如下: ![853bd5f82adfa7755061bd5023c2962](https://user-images.githubusercontent.com/35001756/231431990-330c7430-7b86-42cb-805e-362ae2dda2ab.png) ![28a992cdab3c6885c05a12f25d0099a](https://user-images.githubusercontent.com/35001756/231432009-34003a82-9ce6-4951-9850-1c49f668c173.png)

带cuda的镜像14G,不带cuda的镜像5.8G。 有没有精简的镜像?只需要支持加载模型运行的最简镜像。

deploy

各位大佬,我想请教下,我想解决旋转不变性的角度适应性问题,就是各个角度拍照都能成功的识别。想用方向分类器解决,请问paddleserving有没有提供部署方向分类器的接口?谢谢!

pdServing是否支持自定义的Model? 还是只支持 模型库 https://github.com/PaddlePaddle/Serving/blob/v0.9.0/doc/Model_Zoo_CN.md? 怎样开发来支持自定义的Model

模型串并联问题,如下图,先目标检测,然后根据检测结果再对图片进行分割,分割之后进行图像分类;然后和另一个目标检测的结果进行结果组合。最后一个OP不进行模型预测,只进行前面input_ops的预测结果进行整合。 ![image](https://user-images.githubusercontent.com/36287503/220563803-8d9b2515-53c8-4815-acb5-9e1e13c4cf3d.png) 启动报错 ![da715c481a09f476538da87fca49744](https://user-images.githubusercontent.com/36287503/220565010-f408e2f5-455c-4f09-8126-bb18bea0d556.png) 组合模型的代码 ``` class CombineOp(Op): def preprocess(self, input_data, data_id, log_id): out_data = {} for op_name, data in input_data.items(): if 'bbox_result' in data.keys(): _LOGGER.info("{}: {}".format(op_name, data['bbox_result'])) out_data[op_name] =...