[Bug] The ConvNext Module when used with multiple GPUs causes issues

[rvarun7777]

The bug is due to permute/transpose ops by the layer norm's default change in dims (NCHW) -> (NHWC) which requires these permute ops.

1. The weights are not appropriately loaded for LayerNorm throws a warning only for ConvNext's Image Net 1k Tiny.
2. Due to Layer norm's NCHW -> NHWC ops, there is a permute/transpose op going on. This messes up the distributed training backprop to throw off an error. It works fine with a single GPU.

Python: 3.8.13 [GCC 7.5.0] CUDA available: True GPU 0: Quadro RTX 5000 GPU 1: Quadro RTX 5000 GPU 2: Quadro RTX 5000 CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 11.3, V11.3.109 GCC: gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0 PyTorch: 1.12.1 PyTorch compiling details: PyTorch built with:

• GCC 9.3
• C++ Version: 201402
• Intel(R) oneAPI Math Kernel Library Version 2021.4-Product Build 20210904 for Intel(R) 64 architecture applications
• Intel(R) MKL-DNN v2.6.0 (Git Hash 52b5f107dd9cf10910aaa19cb47f3abf9b349815)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• LAPACK is enabled (usually provided by MKL)
• NNPACK is enabled
• CPU capability usage: AVX2
• CUDA Runtime 11.3
• 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_61,code=sm_61;-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;-gencode;arch=compute_37,code=compute_37
• CuDNN 8.3.2 (built against CUDA 11.5)
• Magma 2.5.2
• Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.3, CUDNN_VERSION=8.3.2, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -fabi-version=11 -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -DEDGE_PROFILER_USE_KINETO -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -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 -Werror=cast-function-type -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.12.1, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=OFF, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF,

TorchVision: 0.13.1 OpenCV: 4.6.0 MMCV: 1.6.0 MMCV Compiler: GCC 9.3 MMCV CUDA Compiler: 11.3 MMDetection: 2.25.1 MMSegmentation: 0.27.0 MMDetection3D: 1.0.0rc3+37aebb8 spconv2.0: True

[tonysy]

Hi, please provide more details of the bug, like the command, environment info, etc.

[rvarun7777]

@tonysy Added and improved the explanation of the issue. It looks like the Pytorch version of ConvNExt had these issues as well. https://github.com/rwightman/pytorch-image-models/pull/1212 Quick fix can be to use the .contiguous() op. This did not work, though!

[JosonChan1998]

@rvarun7777 Hi，I use the .contiguous() op in layernorm2d(). It works to me.

[tonysy]

This issue will be closed as it is inactive, feel free to re-open it if necessary.