q yao comments

Results 318 comments of


                                            q yao

#assertionD:\MMbushu\mmdeploy\csrc\mmdeploy\backend_ops\tensorrt\batched_nms\trt_batched_nms.cpp,103 2022-07-18 11:34:14,224 - mmdeploy - ERROR - visualize tensorrt model failed.

Have you done any modify to the model config?

#assertionD:\MMbushu\mmdeploy\csrc\mmdeploy\backend_ops\tensorrt\batched_nms\trt_batched_nms.cpp,103 2022-07-18 11:34:14,224 - mmdeploy - ERROR - visualize tensorrt model failed.

Could you share your environment? You can use [this script](https://github.com/open-mmlab/mmdeploy/blob/master/tools/check_env.py).

torch2onnx failed for YOLOv3 on Windows

You might need to update the input shape range in the config, for example: ```python backend_config = dict( common_config=dict(max_workspace_size=1

torch2onnx failed for YOLOv3 on Windows

I can not reproduce the error. Could you print the `graph.inputs()`?

torch2onnx failed for YOLOv3 on Windows

https://github.com/onnx/onnx/issues/3493 Try import onnx at the first line in https://github.com/open-mmlab/mmdeploy/blob/14b2bfd524d197565621fedbc47b5940bb22ceed/mmdeploy/apis/onnx/export.py#L122

ERROR: Failed building wheel for onnx

Could you try to locate what caused `check_env.py` crash?

mmdeploy - ERROR - Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

Please post your environment with https://github.com/open-mmlab/mmdeploy/blob/master/tools/check_env.py.

mmdeploy - ERROR - Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

I can not reproduce the error. Please try set the input device and model device in https://github.com/open-mmlab/mmdeploy/blob/master/mmdeploy/apis/onnx/export.py.

How/where to get started developing support for models.

A good startup is trying to add support to part of the model. For example, `yolact` can be split into `backbone`, `neck`, `bbox_head`. Try to convert it with: ```python from...

faster rcnn model deployed to tensorrt with fp32 and the map drop a little

Small mAP loss is inevitable, MMDetection will filter bboxes by scores in RPN, which is hard to implement in TensorRT. I have test the faster-rcnn with DCN and mDCN, both...