[Bug] bad results of RTMW model when the image size is 384x288 #3230
Description
Prerequisite
- I have searched Issues and Discussions but cannot get the expected help.
- The bug has not been fixed in the latest version(https://github.com/open-mmlab/mmpose).
Environment
OrderedDict([('sys.platform', 'linux'), ('Python', '3.11.3 (main, May 15 2023, 15:45:52) [GCC 11.2.0]'), ('CUDA available', True), ('MUSA available', False), ('numpy_random_seed', 2147483648), ('GPU 0', 'Tesla V100-PCIE-16GB'), ('CUDA_HOME', '/usr/local/cuda'), ('NVCC', 'Cuda compilation tools, release 11.7, V11.7.99'), ('GCC', 'gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0'), ('PyTorch', '2.0.0+cu117'), ('PyTorch compiling details', 'PyTorch built with:\n - GCC 9.3\n - C++ Version: 201703\n - Intel(R) oneAPI Math Kernel Library Version 2022.2-Product Build 20220804 for Intel(R) 64 architecture applications\n - Intel(R) MKL-DNN v2.7.3 (Git Hash 6dbeffbae1f23cbbeae17adb7b5b13f1f37c080e)\n - OpenMP 201511 (a.k.a. OpenMP 4.5)\n - LAPACK is enabled (usually provided by MKL)\n - NNPACK is enabled\n - CPU capability usage: AVX2\n - CUDA Runtime 11.7\n - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86\n - CuDNN 8.5\n - Magma 2.6.1\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.7, CUDNN_VERSION=8.5.0, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=0 -fabi-version=11 -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wunused-local-typedefs -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Werror=cast-function-type -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_DISABLE_GPU_ASSERTS=ON, TORCH_VERSION=2.0.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, \n'), ('TorchVision', '0.15.0+cu117'), ('OpenCV', '4.10.0'), ('MMEngine', '0.10.7'), ('MMPose', '1.3.2+71ec36e')])
mmcv 2.1.0
mmdet 3.3.0
mmengine 0.10.7
mmpose 1.3.2
Reproduces the problem - code sample
using topdown_demo_with_mmdet.py file
Reproduces the problem - command or script
x-384:
python demo/topdown_demo_with_mmdet.py
demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py
weights/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth
configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw-x_8xb320-270e_cocktail14-384x288.py
weights/rtmw/rtmw-x_simcc-cocktail14_pt-ucoco_270e-384x288-f840f204_20231122.pth
--input tests/data/coco/git.png
--output-root ./ --draw-bbox
x-256:
python demo/topdown_demo_with_mmdet.py
demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py
weights/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth
configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw-x_8xb704-270e_cocktail14-256x192.py
weights/rtmw/rtmw-x_simcc-cocktail14_pt-ucoco_270e-256x192-13a2546d_20231208.pth
--input tests/data/coco/git.png
--output-root ./ --draw-bbox
l-384:
python demo/topdown_demo_with_mmdet.py
demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py
weights/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth
configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw-l_8xb320-270e_cocktail14-384x288.py
weights/rtmw/rtmw-dw-x-l_simcc-cocktail14_270e-384x288-20231122.pth
--input tests/data/coco/git.png
--output-root ./ --draw-bbox
l-256:
python demo/topdown_demo_with_mmdet.py
demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py
weights/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth
configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw-l_8xb1024-270e_cocktail14-256x192.py
weights/rtmw/rtmw-dw-x-l_simcc-cocktail14_270e-256x192-20231122.pth
--input tests/data/coco/git.png
--output-root ./ --draw-bbox
m-256:
python demo/topdown_demo_with_mmdet.py
demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py
weights/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth
configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw-m_8xb1024-270e_cocktail14-256x192.py
weights/rtmw/rtmw-dw-l-m_simcc-cocktail14_270e-256x192-20231122.pth
--input tests/data/coco/git.png
--output-root ./ --draw-bbox
Reproduces the problem - error message
got original image from : #2741
x-384:
x-256:
l-384:
l-256:
m-256:
Additional information
- I got all weights from https://github.com/open-mmlab/mmpose/blob/main/configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw_cocktail14.md
- When the size is "384x288", the results are bad. And the effect of "rtmw-x_simcc-cocktail14_pt-ucoco_270e-256x192-13a2546d_20231208.pth" model is not so good;
- I tried other images, the conclusion was same;
- I tried different “--kpt-thr” in "topdown_demo_with_mmdet.py" file, didn't help, hard to find the balance;
- Are the provided weights wrong or I should set some params?
Thank you!