train issue with default parameter.

[lji72]

Current learning rate: 0.000100
Current BN decay momentum: 0.500000
2021-03-10 15:41:57.480427
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [31,0,0], thread: [104,0,0] Assertion index Value >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [31,0,0], thread: [456,0,0] Assertion index Value >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [31,0,0], thread: [184,0,0] Assertion index Value >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [31,0,0], thread: [92,0,0] Assertion indexV alue >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [30,0,0], thread: [353,0,0] Assertion index Value >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [30,0,0], thread: [65,0,0] Assertion indexV alue >= 0 && indexValue < src.sizes[dim] failed.
/tmp/pip-req-build-pb3z3zl3/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexTyp
e>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 2]: block: [30,0,0], thread: [9,0,0] Assertion indexVa lue >= 0 && indexValue < src.sizes[dim] failed.

can you help to analyze the reason?
Many thanks！