Lyu Han

> > > > 单独做一个分支吗？ It has been merged into dev-1.x

Please post the result of `python tools/check_env.py` FYI, I can successfully deploy [rotated-faster-rcnn](https://github.com/open-mmlab/mmrotate/tree/dev-1.x/configs/rotated_faster_rcnn) model by the following command: ```bash python tools/deploy.py \ configs/mmrotate/rotated-detection_tensorrt_dynamic-320x320-1024x1024.py /data1/checkpoint/mmrotate/rotated-faster-rcnn-le90_r50_fpn_1x_dota.py \ /data1/checkpoint/mmrotate/rotated_faster_rcnn_r50_fpn_1x_dota_le90-0393aa5c.pth \ ../mmrotate/demo/demo.jpg \ --dump-info...

Based on FasterTransformer, we have implemented an efficient inference engine - [TurboMind](https://github.com/InternLM/lmdeploy#introduction), supporting both llama and llama-2

Hi, thanks for your attention to mmdeploy. We haven't planned it yet. Would you like to give it a try? `mmdeploy/codebase/mmcls/*` can be reviewed as an example.

What's the converted result? You can find `output_tensorrt.jpg` under the working directory that is passed to `tools/deploy.py`.

Hope the guide in [this ](https://github.com/OpenGVLab/InternVL/pull/152) PR can do a favor

If you `echo $LD_LIBRARY_PATH` in another shell, what will print? Have you tried to reopen PyCharm?

@irexyc would you please help checking this issue on Windows platform?

Try to use this deployment config `detection_tensorrt_dynamic-64x64-608x608.py`, since the maximum input resolution of yolov3 from mmdet is `img_scale=(320, 320)`

权重量化为4 bit后，应该可以。[lmdeploy](https://github.com/InternLM/lmdeploy) 正在开发相关的功能，欢迎关注 :smile: