KeyError: "BasicVSR: 'BasicVSRPlusPlus is not in the model registry'"

[NasrinR791]

Hi,

I am training BasicVSR++ from scratch on HPC cluster with following environment info:

`> sys.platform: linux
Python: 3.8.12 (default, Oct 12 2021, 13:49:34) [GCC 7.5.0]
CUDA available: True
GPU 0: Tesla T4
CUDA_HOME: /usr/local/cuda
NVCC: Build cuda_11.0_bu.TC445_37.28845127_0
GCC: gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-16)
PyTorch: 1.9.0
PyTorch compiling details: PyTorch built with:

• GCC 7.3
• 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 v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• NNPACK is enabled
• CPU capability usage: AVX2
• CUDA Runtime 10.2
• 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_37,code=compute_37
• CuDNN 7.6.5
• Magma 2.5.2
• Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=10.2, CUDNN_VERSION=7.6.5, 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=1.9.0, 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=ON, USE_NNPACK=ON, USE_OPENMP=ON,

TorchVision: 0.10.0
OpenCV: 4.5.4
MMCV: 1.4.6
MMCV Compiler: GCC 7.3
MMCV CUDA Compiler: 10.2
MMEditing: 0.13.0+793d026
`

After stopping the training due to time limit in cluster, I resume the training using config file basicvsr_plusplus_c64n7_8x1_600k_reds4.py :
--resume-from .../results/BasicVSR_plus_plus/train/iter_565000.pth
However I got the following error:
KeyError: "BasicVSR: 'BasicVSRPlusPlus is not in the model registry'"

This is while I resumed the training two times ago and it worked without problem.

The complete error message:
`Traceback (most recent call last):
File "/kuacc/users/nrahimi19/miniconda3/envs/mmlab/lib/python3.8/site-packages/mmcv/utils/registry.py", line 52, in build_from_cfg
return obj_cls(**args)
File "/scratch/users/nrahimi19/mmediting/mmedit/models/restorers/basicvsr.py", line 40, in init
super().init(generator, pixel_loss, train_cfg, test_cfg,
File "/scratch/users/nrahimi19/mmediting/mmedit/models/restorers/basic_restorer.py", line 48, in init
self.generator = build_backbone(generator)
File "/scratch/users/nrahimi19/mmediting/mmedit/models/builder.py", line 31, in build_backbone
return build(cfg, BACKBONES)
File "/scratch/users/nrahimi19/mmediting/mmedit/models/builder.py", line 22, in build
return build_from_cfg(cfg, registry, default_args)
File "/kuacc/users/nrahimi19/miniconda3/envs/mmlab/lib/python3.8/site-packages/mmcv/utils/registry.py", line 44, in build_from_cfg
raise KeyError(
KeyError: 'BasicVSRPlusPlus is not in the model registry'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/scratch/users/nrahimi19/mmediting/tools/train.py", line 150, in
main()
File "/scratch/users/nrahimi19/mmediting/tools/train.py", line 113, in main
model = build_model(
File "/scratch/users/nrahimi19/mmediting/mmedit/models/builder.py", line 60, in build_model
return build(cfg, MODELS, dict(train_cfg=train_cfg, test_cfg=test_cfg))
File "/scratch/users/nrahimi19/mmediting/mmedit/models/builder.py", line 22, in build
return build_from_cfg(cfg, registry, default_args)
File "/kuacc/users/nrahimi19/miniconda3/envs/mmlab/lib/python3.8/site-packages/mmcv/utils/registry.py", line 55, in build_from_cfg
raise type(e)(f'{obj_cls.name}: {e}')
KeyError: "BasicVSR: 'BasicVSRPlusPlus is not in the model registry'"
`

Could you please help?

[ckkelvinchan]

BasicVSR++ should not be in the model registry. It should be in the backbone registry.

[NasrinR791]

Thanks for your response @ckkelvinchan
How can I fix it? Using config file?
I did change nothing in config file but --resume-from .../results/BasicVSR_plus_plus/train/iter_565000.pth. What may cause this error?

[ckkelvinchan]

This is a bit strange. How about trying to use the resume_from in the configuration file?

[NasrinR791]

Dear @ckkelvinchan

I found out that I have removed @BACKBONES.register_module() in basicvsr_pp.py file by mistake,
I fix it and it works without problem.
Thank you