[Bug] SoftmaxFocalLossFunction backward bug

[kaiyi98]

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/mmcv).

Environment

{'sys.platform': 'linux', 'Python': '3.8.15 (default, Nov 4 2022, 20:59:55) [GCC 11.2.0]', 'CUDA available': True, 'GPU 0,1': 'Tesla V100-PCIE-16GB', 'CUDA_HOME': '/usr/local/cuda', 'NVCC': 'Build cuda_11.3.r11.3/compiler.29920130_0', 'GCC': 'gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0', 'PyTorch': '1.9.1+cu111', 'PyTorch compiling details': 'PyTorch built with:\n - GCC 7.3\n - C++ Version: 201402\n - Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications\n - Intel(R) MKL-DNN v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb)\n - OpenMP 201511 (a.k.a. OpenMP 4.5)\n - NNPACK is enabled\n - CPU capability usage: AVX2\n - CUDA Runtime 11.1\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 8.0.5\n - Magma 2.5.2\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -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 -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.9.1, 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=ON, USE_NNPACK=ON, USE_OPENMP=ON, \n', 'TorchVision': '0.10.1+cu111', 'OpenCV': '4.9.0', 'MMCV': '1.4.0', 'MMCV Compiler': 'GCC 7.5', 'MMCV CUDA Compiler': '11.3'}

Reproduces the problem - code sample

def backward(ctx, grad_output):
input_softmax, target, weight = ctx.saved_tensors
buff = input_softmax.new_zeros(input_softmax.size(0))
grad_input = input_softmax.new_zeros(input_softmax.size())
print('input_softmax ', input_softmax.shape)
ext_module.softmax_focal_loss_backward(
input_softmax,
target,
weight,
buff,
grad_input,
gamma=ctx.gamma,
alpha=ctx.alpha)
print('grad_output ',grad_output.shape)
grad_input *= grad_output

Reproduces the problem - command or script

that's a train model project.

Reproduces the problem - error message

RuntimeError: The size of tensor a (3) must match the size of tensor b (150) at non-singleton dimension 1

Additional information

I want to use softmax_focal_loss to calculate classification loss. When it forward, the prediction shape is [n, 3](n is the sample number), and the target shape is [n]. The forward calculate the true loss，but when it backward, the error occurred.