THCudaCheck FAIL file=/opt/conda/conda-bld/pytorch_1556653099582/work/aten/src/THCUNN/generic/LeakyReLU.cu line=29 error=48 : no kernel image is available for execution on the device

[Youthfeng123]

I follow the instructions on INSTALL.md, and install the packgage with no error report. howerver, when I try to run the following code:
python scripts/demo_inference.py --cfg configs/coco/resnet/256x192_res50_lr1e-3_1x.yaml --checkpoint pretrained_models/fast_res50_256x192.pth --indir examples/demo/ --sp

It stucked here for a long time:
Loading YOLO model..
Loading pose model from pretrained_models/fast_res50_256x192.pth...

after tens of minuts, it report this error:
THCudaCheck FAIL file=/opt/conda/conda-bld/pytorch_1556653099582/work/aten/src/THCUNN/generic/LeakyReLU.cu line=29 error=48 : no kernel image is available for execution on the device
Exception in thread Thread-2:
Traceback (most recent call last):
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/threading.py", line 864, in run
self._target(*self._args, **self._kwargs)
File "/home/xdlab3/DeepTech/AlphaPose-master/alphapose/utils/detector.py", line 223, in image_detection
dets = self.detector.images_detection(imgs, im_dim_list)
File "/home/xdlab3/DeepTech/AlphaPose-master/detector/yolo_api.py", line 94, in images_detection
prediction = self.model(imgs, args=args)
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/home/xdlab3/DeepTech/AlphaPose-master/detector/yolo/darknet.py", line 332, in forward
x = self.module_listi
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/modules/container.py", line 92, in forward
input = module(input)
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/modules/activation.py", line 561, in forward
return F.leaky_relu(input, self.negative_slope, self.inplace)
File "/home/xdlab3/anaconda3/envs/alphapose/lib/python3.6/site-packages/torch/nn/functional.py", line 1097, in leaky_relu
result = torch._C.nn.leaky_relu(input, negative_slope)
RuntimeError: cuda runtime error (48) : no kernel image is available for execution on the device at /opt/conda/conda-bld/pytorch_1556653099582/work/aten/src/THCUNN/generic/LeakyReLU.cu:29

0%| | 0/3 [00:00<?, ?it/s] 0%| | 0/3 [83:17:25<?, ?it/s]
===========================> Finish Model Running.
Results have been written to json.

I use 3090, with pytorch==1.1.0.
and cuda is 11.1.
though the version may be incompatible, I ran torch.cuda.is_available(), it returned True.
and I can put tensors on GPU at will.
So how would I fix this? thanks a lot!

[peteruhrig]

Can you try this pull request?
#1032
It works with A40 and A100; RTX 3090 is the same generation, so I think it should work.

[Fang-Haoshu]

Thanks @peteruhrig , we have merged it to the master branch