我用PaddleOCR训练出了一个识别模型，但是沙盒部署没有预测结果，

日志：

{
"logId": "911e8fce-840b-47e5-918f-886639667619",
"errorCode": 0,
"errorMsg": "成功",
"timestamp": 0,
"result": "start deploy...\nget file url from evn\ndownloading...\nextracting...\nunzip is /usr/bin/unzip\nunzip available\nsource hook dir:webserver/home/hook/inputhook.py\nhook ok\nsource model dir:webserver/model/work/PaddleOCR/inference/rec/bs/inference.pdmodel\nmodel ok\nstart server...\n[2022-05-06 07:*:*,850.850] server_conf.py [line:*] INFO current config section cpu\n[2022-05-06 07:*:*,405.405] executor_manager.py [line:*] INFO device type:cpu executor count:* model dir:{'model': '/home/work/serving/resources/model/'} get executor timeout:*.5s\n[2022-05-06 07:*:*,406.406] app_server.py [line:*] INFO start app listen port:* thread pool size:* site root:serving\n[2022-05-06 07:*:*,648.648] infer_service.py [line:*] WARNING infer param not match model requirement! require params: ['img']\n[2022-05-06 07:*:*,648.648] business_service.py [line:*] ERROR infer exception: infer param not match model requirement! require params: ['img']\n[2022-05-06 07:*:*,648.648] app_server.py [line:*] INFO request cost:*.05 sec\n"
}

参数转换器如下图：