DerryHub/BEVFormer_tensorrt issues and pull requests

#101 - why overwrite multiscaleDeformableAttnPlugin ?

Issue - State: open - Opened by lix19937 over 1 year ago

#100 - UNSUPPORTED_NODE: Assertion failed: creator && "Plugin not found, a When running test sample

Issue - State: open - Opened by guiyuliu over 1 year ago - 2 comments

#99 - About onnx surgeon

Issue - State: closed - Opened by gaosanyuan over 1 year ago - 2 comments

#98 - AttributeError: 'NoneType' object has no attribute 'set_binding_shape'

Issue - State: open - Opened by Alex-fishred over 1 year ago

#97 - No importer registered for op: MMCVModulatedDeformConv2d. Attempting to import as plugin.

Issue - State: open - Opened by Alex-fishred over 1 year ago - 2 comments

#96 - fatal error: cudnn.h: No such file or directory

Issue - State: open - Opened by Alex-fishred over 1 year ago - 2 comments

#95 - why tensorrt do inference slower than pytorch

Issue - State: open - Opened by gofreelee over 1 year ago

#94 - Question about the fp16 mode of mmdeploy grid_sampler

Issue - State: open - Opened by kamiLight over 1 year ago - 1 comment

#93 - Are the Inverse, BEV Pool V2, Flash Multi-Head Attention not used in the bevformer?

Issue - State: open - Opened by kamiLight over 1 year ago

#92 - identifier is undefined when include <cuda_fp16.h>

Issue - State: open - Opened by kamiLight over 1 year ago

#91 - Did you really run the BEVFromer(Pytorch) on 3090? Or you just copy the result from https://github.com/fundamentalvision/BEVFormer

Issue - State: open - Opened by admyxs over 1 year ago

#90 - OSError: /home/work/Downloads/BEVFormer_tensorrt/TensorRT/lib/libtensorrt_ops.so: cannot open shared object file: No such file or directory

Issue - State: open - Opened by ZSA010807 over 1 year ago - 3 comments

#89 - What's the purpose of prod for bev_mask

Issue - State: open - Opened by Jian-danai over 1 year ago

#88 - RuntimeError: modulated_deformable_im2col_impl: implementation for device cuda:0 not found.

Issue - State: open - Opened by qingwan7 over 1 year ago - 2 comments

#87 - Update README.md

Pull Request - State: closed - Opened by DerryHub over 1 year ago

#86 - TensorRT plugin for BEVFormer Encoder

Issue - State: closed - Opened by Wolfybox over 1 year ago - 1 comment

#85 - The input image size

Issue - State: closed - Opened by Jian-danai over 1 year ago - 2 comments

#84 - 'CustomMSDeformableAttentionTRTP' gives incorrect outputs in Pytorch inference

Issue - State: closed - Opened by Jian-danai over 1 year ago - 2 comments

#83 - Stucked at onnx::Expand_1007 when running "sh samples/bevformer/plugin/tiny/onnx2trt.sh -d 0"

Issue - State: open - Opened by kamiLight over 1 year ago - 1 comment

#82 - Update README.md

Pull Request - State: closed - Opened by DerryHub over 1 year ago

#81 - add softmax into MSDA & speed up TRT OPs

Pull Request - State: closed - Opened by DerryHub over 1 year ago

#80 - ERROR: Could not build wheels for pycocotools, which is required to install pyproject.toml-based projects

Issue - State: open - Opened by kamiLight over 1 year ago - 3 comments

#79 - A bug when transferring onnx to trt

Issue - State: open - Opened by Son-Goku-gpu over 1 year ago - 1 comment

#78 - Is there a docker tar package that can be used directly？

Issue - State: open - Opened by qingwan7 almost 2 years ago - 9 comments

#77 - Dockerfile half updated?

Issue - State: open - Opened by jeeprhymeallthetime almost 2 years ago - 1 comment

#76 - Docker install and mmlabs fail on git clone

Issue - State: closed - Opened by jeeprhymeallthetime almost 2 years ago - 1 comment

#75 - benchmark on orin

Issue - State: open - Opened by shuyuan-wang almost 2 years ago - 4 comments

#74 - amples/bevformer/tiny/pth2onnx.sh error,

Issue - State: open - Opened by linClubs almost 2 years ago - 2 comments

#73 - qat model

Issue - State: open - Opened by CodeWorldChanged almost 2 years ago - 1 comment

#72 - Can't find BEVFormer_ptq_max model

Issue - State: closed - Opened by Oh-Cest-la-vie almost 2 years ago - 2 comments

#71 - SpatialCrossAttentionTRT don't select valid query, no cuda memory problem here?

Issue - State: closed - Opened by sherylwang almost 2 years ago - 18 comments

#70 - error: no instance of overloaded function "nvcuda::wmma::load_matrix_sync"

Issue - State: closed - Opened by santoshcoder23 almost 2 years ago - 5 comments

#69 - error in ms_deformable_im2cal_cuda : no kernel image is available for execution on the device

Issue - State: closed - Opened by CodeWorldChanged almost 2 years ago - 21 comments

#68 - what's the difference between mmdeploy plugins and custom plugins ?

Issue - State: closed - Opened by reverse-2020 almost 2 years ago - 6 comments

#67 - Update multi_head_attn.py

Pull Request - State: closed - Opened by DerryHub about 2 years ago

#66 - The usage of flash attention

Issue - State: closed - Opened by 2401ch about 2 years ago - 2 comments

#65 - When doing yolox/trt_evaluate.sh for fp32 tensorrt model, what's the reason this error occurred? how can i fix it?

Issue - State: open - Opened by NutshellLee about 2 years ago - 2 comments

#64 - How to use a different yolox weight other than the one provided to do quantization in the context?

Issue - State: open - Opened by NutshellLee about 2 years ago - 1 comment

#63 - What does points_per_group mean？

Issue - State: open - Opened by shaoqb about 2 years ago - 2 comments

#62 - add Dockerfile and requirements.txt

Pull Request - State: closed - Opened by DerryHub about 2 years ago

#61 - No module named 'pytorch_quantization

Issue - State: closed - Opened by Isaac-xie about 2 years ago - 1 comment

#60 - support Flash Multi-Head Attention Plugin (FP32/FP16/INT8)

Pull Request - State: closed - Opened by DerryHub about 2 years ago

#59 - ScatterND error when running onnx2trt.sh with TensorRT 8.5.2

Issue - State: open - Opened by sun-lingyu about 2 years ago - 1 comment

#58 - support Deployment of BEVDet

Pull Request - State: closed - Opened by DerryHub about 2 years ago

#57 - does anyone try to deploy this great repo in jetson orin? Unsupported operator Gridsampler2DTRT

Issue - State: open - Opened by sainttelant about 2 years ago - 6 comments

#56 - How can i see the output image

Issue - State: open - Opened by Ashu683 about 2 years ago

#55 - build bev_mmdert3d error: Invalid distribution name or version syntax: bev-mmdet3d--0.1

Issue - State: closed - Opened by firestonelib about 2 years ago - 1 comment

#54 - fix Invalid distribution name or version syntax: bev-mmdet3d--0.1 bug

Pull Request - State: closed - Opened by firestonelib about 2 years ago

#53 - ImportError: cannot import name 'iou3d_cuda'

Issue - State: closed - Opened by Ashu683 about 2 years ago

#52 - ImportError: cannot import name 'iou3d_cuda' from partially initialized module 'third_party.bev_mmdet3d.ops.iou3d'

Issue - State: closed - Opened by Ashu683 about 2 years ago - 1 comment

#51 - can not use onnxsim

Issue - State: open - Opened by shaoqb about 2 years ago - 3 comments

#50 - Any plans for bevfusion TRT

Issue - State: open - Opened by lucianzhong over 2 years ago - 2 comments

#49 - Rename BEV Package

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#48 - how can i visualize the evaluation results?

Issue - State: open - Opened by sainttelant over 2 years ago

#47 - Error executing samples/test_trt_ops.sh

Issue - State: closed - Opened by sun-lingyu over 2 years ago - 4 comments

#46 - deserialize_cuda_engine(): incompatible function arguments

Issue - State: closed - Opened by shaoqb over 2 years ago - 5 comments

#45 - No importer registered for op: nan_to_num. Attempting to import as plugin.

Issue - State: open - Opened by xll426 over 2 years ago - 2 comments

#44 - INT8 accuray issue for Multiscale deformable attention

Issue - State: open - Opened by pianogGG over 2 years ago - 3 comments

#43 - Dockerfile for installing the project.

Issue - State: closed - Opened by matcosta23 over 2 years ago - 2 comments

#42 - the shape inference of mmcv::MMCVModulatedDeformConv2d type is missing

Issue - State: closed - Opened by TGpastor over 2 years ago - 14 comments

#41 - modified README.md

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#40 - trt evaluate error

Issue - State: open - Opened by HerrYu123 over 2 years ago - 3 comments

#39 - TypeError: deserialize_cuda_engine(): incompatible function arguments. The following argument types are supported: 1. (self: tensorrt.tensorrt.Runtime, serialized_engine: buffer) -> tensorrt.tensorrt.ICudaEngine

Issue - State: open - Opened by Tanzichang over 2 years ago - 10 comments

#38 - fix error: expression must be a modifiable lvalue

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#37 - onnx2trt with nv_half and nv_half2 failed

Issue - State: open - Opened by WYYAHYT over 2 years ago - 9 comments

#36 - model download link is not work

Issue - State: closed - Opened by qiaosiyi over 2 years ago - 1 comment

#35 - Install project error

Issue - State: closed - Opened by HerrYu123 over 2 years ago - 10 comments

#34 - modified test_grid_sampler.py

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#33 - support FP16/INT8 to DCN Plugin

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#32 - support FP16/INT8 to Rotate Plugin

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#31 - update README.md

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#30 - modified test_multi_scale_deformable_attn.py

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#29 - fix T2int8 to round-to-nearest-even

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#28 - fix cublas error of DCN plugin: CUBLAS_STATUS_NOT_SUPPORTED

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#27 - onnx2trt failed with UNSUPPORTED_NODE

Issue - State: closed - Opened by WYYAHYT over 2 years ago - 11 comments

#26 - torch.onnx.export transform bevformer to onnx occure error aten::mul not support

Issue - State: closed - Opened by zxros10 over 2 years ago - 1 comment

#25 - Reduce precision error of MSDA in INT8 mode

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#24 - Why pth_evaluate.sh use bevformer_base_trt.py config

Issue - State: closed - Opened by WYYAHYT over 2 years ago - 2 comments

#23 - support INT8 inference to MultiScaleDeformableAttn

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#22 - could you please share BEFormer_tiny onnx model?

Issue - State: closed - Opened by RunnerZhong over 2 years ago - 4 comments

#21 - Process finished with exit code 136 (interrupted by signal 8: SIGFPE)

Issue - State: closed - Opened by lfs119 over 2 years ago - 7 comments

#20 - support INT8 inference to rotate

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#19 - fix dcn plugin configure

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#18 - support INT8 inference to DCN

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#17 - support INT8 inference to GridSampler2D

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#16 - sh sample/test_trt_ops.sh error

Issue - State: closed - Opened by zouzouwei over 2 years ago - 3 comments

#15 - support kCHW2 Tensor Format to Rotate

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#14 - support kCHW2 Tensor Format to ModulatedDeformableConv2d

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#13 - support kCHW2 Tensor Format to GridSampler

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#12 - format code

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#11 - Simplify BEVFormer forward

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#10 - IndexError: Argument passed to at() was not in the map.

Issue - State: closed - Opened by pratiklikharOLA over 2 years ago - 6 comments

#9 - missing template specialization for type<__half> of function template "ms_deformable_im2col_cuda"

Issue - State: closed - Opened by hygxy over 2 years ago - 1 comment

#8 - Support TensorRT Calibrators

Pull Request - State: closed - Opened by DerryHub over 2 years ago

#7 - TypeError: deserialize_cuda_engine(): incompatible function arguments. The following argument types are supported: 1. (self: tensorrt.tensorrt.Runtime, serialized_engine: buffer) -> tensorrt.tensorrt.ICudaEngine

Issue - State: closed - Opened by chaojiyu over 2 years ago - 7 comments

#6 - AssertionError: Samples in split doesn't match samples in predictions.

Issue - State: closed - Opened by hbc2020 over 2 years ago - 3 comments

#5 - RuntimeError: CUDA out of memory. Tried to allocate 1.91 GiB (GPU 0; 11.77 GiB total capacity; 8.90 GiB already allocated; 509.19 MiB free; 8.90 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

Issue - State: closed - Opened by yhwang-hub over 2 years ago - 1 comment

GitHub / DerryHub/BEVFormer_tensorrt issues and pull requests