BEVFormer_tensorrt
BEVFormer_tensorrt copied to clipboard
BEVFormer inference on TensorRT, including INT8 Quantization and Custom TensorRT Plugins (float/half/half2/int8).
I'm trying to Build and Install Custom TensorRT Plugins When I execute make -j$(nproc) fatal error: cudnn.h: No such file or directory #include ^ compilation terminated. make[2]: *** [CMakeFiles/tensorrt_ops.dir/build.make:206: CMakeFiles/tensorrt_ops.dir/plugin/multi_head_attn/multiHeadAttnPlugin.cpp.o]...
I am executing onnx2trt.sh but got error How to solve it? 
I have the following problems when I run"sh samples/test_trt_ops.sh", how can I solve them? OSError: /home/work/Downloads/BEVFormer_tensorrt/TensorRT/lib/libtensorrt_ops.so: cannot open shared object file: No such file or directory
Device: NVIDIA A100 80G tensorrt fps: 2.61 pytorch fps: 2.94 Why is tensorrt's performance even lower
Thanks for your great work! I found the Grid Sampler, Multi-scale Deformable Attention, Modulated Deformable Conv2d, Rotate plugins are used in the bevformer, but i couldn't find others, i.e. Inverse,...
when include in .cu, it will report the identifier xxx is undefined like this: ``` rotateKernel.cu(40): error: identifier "__hdiv" is undefined ``` GPU: RTX 3090 system: ubuntu 18.04
Did you really run the BEVFromer(Pytorch) on 3090? Or you just copy the result from https://github.com/fundamentalvision/BEVFormer. I found the bevformer_tiny runned on 8 A100-SXM-80GB in https://github.com/fundamentalvision/BEVFormer, you run bevformer_tiny on...
https://github.com/DerryHub/BEVFormer_tensorrt/blob/303d3140c14016047c07f9db73312af364f0dd7c/det2trt/models/modules/encoder.py#L256C10-L258C84 ``` bev_mask = (1 - (1 - bev_mask).prod(0)).view(6, -1, 1) bev_mask = bev_mask / torch.clamp(bev_mask.sum(0, keepdims=True), min=1e-4) ``` What's the purpose of this, it makes `bev_mask` different from original...
I have the following problems when I run"sh samples/bevformer/plugin/base/pth2onnx.sh", how can I solve them? RuntimeError: modulated_deformable_im2col_impl: implementation for device cuda:0 not found.