AllentDan

`mask_rcnn` is of instance-seg. Please use instance-seg configs.

Hi, @Phelan164 please refer to [the supported model list](https://github.com/open-mmlab/mmdeploy/blob/master/docs/en/03-benchmark/supported_models.md). Mask RCNN for text detection is not in the list.

Could you run `python tools/check_env.py` to collect necessary environment information and paste it here?

Hi, @Audrey528. There are two methods. 1. The test pipeline of model_cfg can be modified to your preference. 2. You can also modify the function `create_input` of `mmdeploy/codebase/mmdet/deploy/object_detection.py` to change...

Set `keep_ratio=False` please

Hi, could you try debugging `create_input` function of `mmdeploy/codebase/mmdet/deploy/object_detection.py` and check what step is not expected?

Hi, @ultimate1102. Could you please help test dbnet-r18 and check whether the ORT is slower than PyTorch?

Hi, @8633136 @ultimate1102. Did you warm up before obtaining inference time?

Well, our deployment routine follows PyTorch->ONNX->TensorRT. Tips are: 1. build and enjoy MMDeploy first though [get_started](https://github.com/open-mmlab/mmdeploy/blob/master/docs/en/get_started.md). It would be better if you convert an OCR model to TensorRT and visualize...

Hi, @anuj-rathore. SDMG-R is not in our short-term plan. But just like @gaotongxiao said, we welcome that you can give a try on that. Regards.