An error occured when I trained the model on my own dataset.

[Sagiri18]

When I recurrent the R-C3D model to achieve the action recognization task on my own dataset, I had the same problem.

[session 1][epoch  1][iter    1/  30] loss: 2.5127, lr: 1.00e-04
		fg/bg=(5/123), gt_twins: 1, time cost: 53.711857
		rpn_cls: 0.7111, rpn_twin: 0.0049, rcnn_cls: 1.7332, rcnn_twin 0.0635
one step data time: 2.2378
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [0,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [1,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [2,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [3,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [4,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [5,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [6,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [7,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [8,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [9,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [10,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [11,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [12,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [13,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [14,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
/pytorch/aten/src/THC/THCTensorScatterGather.cu:100: void THCudaTensor_gatherKernel(TensorInfo<Real, IndexType>, TensorInfo<Real, IndexType>, TensorInfo<long, IndexType>, int, IndexType) [with IndexType = unsigned int, Real = float, Dims = 3]: block: [0,0,0], thread: [15,0,0] Assertion `indexValue >= 0 && indexValue < src.sizes[dim]` failed.
Traceback (most recent call last):
  File "/content/R-C3D.pytorch-pytorch-1.1/trainval_net.py", line 360, in <module>
train_net(tdcnn_demo, dataloader, optimizer, args)
  File "/content/R-C3D.pytorch-pytorch-1.1/trainval_net.py", line 161, in train_net
loss_temp += reduced_loss.item()
RuntimeError: CUDA error: device-side assert triggered

I‘m a rookie, can anyone give me some advice? Thanks a lot!
The environment is:

1. Python 3.7
2. CUDA 10.0
3. Pytorch 1.1.0
4. Torchvision 0.3.0

By the way, I use Google Colab to do this job. The default version of CUDA is 11.2, but the version I need is 10.0. I noticed that there is CUDA 10.0 under path “usr/loacl” , I pointed the soft connection of "usr/local/CUDA" to "usr/local/CUDA10.0", but When I check the GPU information, it shows CUDA 11.2 which confused me.

[Mary-xl]

any solution to it? Got the same issue here

[Mary-xl]

Hey buddy, check detclasslist.txt to see if it starts from 0. I solved the issue by correcting this part.

[BigBuffa1o]

Hey buddy, check detclasslist.txt to see if it starts from 0. I solved the issue by correcting this part.

i meet same problem by start at 1,should it start at 0 or 1?If start at 0 i assmue backgroud is 0?If start at 1 does that means when set num_classes we should add 1?(if we got 8 classes we should write 9 for config instead)