Error when running pytorch1.9 version

[xjtAlgo]

(rgo-py37-f) fy@neu:~/DCNv2-pytorch_1.9/test$ python test.py
error in modulated_deformable_im2col_cuda: no kernel image is available for execution on the device
error in modulated_deformable_col2im_coord_cuda: no kernel image is available for execution on the device
error in modulated_deformable_col2im_cuda: no kernel image is available for execution on the device
error in modulated_deformable_im2col_cuda: no kernel image is available for execution on the device
error in modulated_deformable_col2im_coord_cuda: no kernel image is available for execution on the device
error in modulated_deformable_col2im_cuda: no kernel image is available for execution on the device
error in modulated_deformable_im2col_cuda: no kernel image is available for execution on the device
torch.Size([2, 64, 128, 128])
THCudaCheck FAIL file=/home/fy/DCNv2-pytorch_1.9/src/cuda/dcn_v2_psroi_pooling_cuda.cu line=339 error=209 : no kernel image is available for execution on the device
Traceback (most recent call last):
File "test.py", line 285, in
example_dpooling()
File "test.py", line 238, in example_dpooling
out = pooling(input, rois, offset)
File "/home/fy/anaconda3/envs/rgo-py37-f/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl
return forward_call(*input, **kwargs)
File "/home/fy/DCNv2-pytorch_1.9/dcn_v2.py", line 278, in forward
self.trans_std,
File "/home/fy/DCNv2-pytorch_1.9/dcn_v2.py", line 210, in forward
ctx.trans_std,
RuntimeError: cuda runtime error (209) : no kernel image is available for execution on the device at /home/fy/DCNv2-pytorch_1.9/src/cuda/dcn_v2_psroi_pooling_cuda.cu:339

I can compile successfully. But when I run the test.py, it will go wrong with deformable pooling.My graphics card is rtx3090, pytorch=1.9.0 CUDA=11.1.
What should i do, could you give me some suggestions？Thanks.

[xjtAlgo]

(rgo-py37) fy@neu:~/DCNv2-pytorch_1.7/test$ python test.py
error in modulated_deformable_im2col_cuda: no kernel image is available for execution on the device
Traceback (most recent call last):
File "test.py", line 284, in
example_dconv()
File "test.py", line 202, in example_dconv
error.backward()
File "/home/fy/anaconda3/envs/rgo-py37/lib/python3.7/site-packages/torch/tensor.py", line 221, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/home/fy/anaconda3/envs/rgo-py37/lib/python3.7/site-packages/torch/autograd/init.py", line 132, in backward
allow_unreachable=True) # allow_unreachable flag
File "/home/fy/anaconda3/envs/rgo-py37/lib/python3.7/site-packages/torch/autograd/function.py", line 89, in apply
return self._forward_cls.backward(self, *args) # type: ignore
File "/home/fy/anaconda3/envs/rgo-py37/lib/python3.7/site-packages/torch/autograd/function.py", line 210, in wrapper
outputs = fn(ctx, *args)
File "/home/fy/DCNv2-pytorch_1.7/dcn_v2.py", line 61, in backward
ctx.deformable_groups,
RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling cublasCreate(handle)

When I run the pytorch 1.7 version of the test.py, it also raise an error.

[xjtAlgo]

The two environments I installed above can train the model, but neither of them can use DCN.