Vijaysrinivas Rajagopal

This could be helpful for anyone looking for something similar: https://github.com/ZhangGe6/onnx-modifier It builds on top of Netron.

Does this help? https://github.com/dbolya/tide/issues/19

Any updates regarding adding your own custom dataset drivers?

If someone wants an example of what @drewm1980 had mentioned to get a custom dataset driver working, here's an example of another non-COCO-related dataset working (VOC2007): https://gist.github.com/vjsrinivas/56ca6e209adf23be17b9d2266b288c71 It's still calculating...

I believe this is just an unfortunate limitation of GPU kernel initialization. There is not much you can do other than maybe running dummy inputs for however many slow inferences...

Experiencing the same problem as @wstnturner. Even building OpenCV from the repo like mentioned earlier failed with this message: ``` [ 85%] Building CXX object modules/ximgproc/CMakeFiles/opencv_ximgproc.dir/opencv_ximgproc_main.cpp.obj [ 85%] Linking CXX...

Has there been any updates regarding ONNX support?

I'm sorry. I don't understand the question. Is the code snippet I sent causing the screenshot you posted?

@freedenS I did try on trt7 and it worked fine. I uploaded the changes to my fork: https://github.com/vjsrinivas/tensorrtx/pull/1/files I'll look into layer changes in trt8 and maybe the output shapes...

Investigated the `createEngine_r50detr` function. Printed out all the tensor dimensions for `features`, `pos_embed`, `input_proj`, `flatten`, and `results`. They are identical between TRT7 and TRT8 machines. I also cut up the...