TX2 cuda runtime error (7) : too many resources requested for launch ，运行错误解决方法

最新推荐文章于 2024-06-29 19:16:36 发布

丶Shining

最新推荐文章于 2024-06-29 19:16:36 发布

阅读量7k

点赞数

分类专栏： pytorch技巧 Linux部署 Hi3559A 技巧方法文章标签： TX2

本文链接：https://blog.csdn.net/xxradon/article/details/87922618

版权

Linux部署同时被 3 个专栏收录

45 篇文章 0 订阅

订阅专栏

技巧方法

17 篇文章 0 订阅

订阅专栏

pytorch技巧

15 篇文章 2 订阅

订阅专栏

请先更新pytorch源码至最新的master，然后重新编译，看是否解决了，如果没有了尝试下面的解决方法。

RuntimeError: cuda runtime error (7) : too many resources requested for launch at /home/nvidia/Downloads/pytorch/aten/src/THCUNN/generic/SpatialUpSamplingBilinear.cu:66

Full log:

CUDA_ENABLED:  True
/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/upsampling.py:225: UserWarning: nn.UpsamplingBilinear2d is deprecated. Use nn.functional.interpolate instead.
  warnings.warn("nn.UpsamplingBilinear2d is deprecated. Use nn.functional.interpolate instead.")
/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/upsampling.py:122: UserWarning: nn.Upsampling is deprecated. Use nn.functional.interpolate instead.
  warnings.warn("nn.Upsampling is deprecated. Use nn.functional.interpolate instead.")
THCudaCheck FAIL file=/home/nvidia/Downloads/pytorch/aten/src/THCUNN/generic/SpatialUpSamplingBilinear.cu line=66 error=7 : too many resources requested forlaunch
Traceback (most recent call last):
  File "train_binseg.py", line 73, in <module>
    outputs = model(inputs)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/nvidia/Downloads/deep-learning-pytorch/segmentation/networks/SegNet.py", line 73, in forward
    enc5 = self.enc5(dec5)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/nvidia/Downloads/deep-learning-pytorch/segmentation/networks/SegNet.py", line 33, in forward
    return self.encode(x)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/container.py", line 91, in forward
    input = module(input)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/upsampling.py", line 226, in forward
    return super(UpsamplingBilinear2d, self).forward(input)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/modules/upsampling.py", line 123, in forward
    return F.interpolate(input, self.size, self.scale_factor, self.mode, self.align_corners)
  File "/home/nvidia/.local/lib/python3.5/site-packages/torch/nn/functional.py", line 1985, in interpolate
    return torch._C._nn.upsample_bilinear2d(input, _output_size(2), align_corners)
RuntimeError: cuda runtime error (7) : too many resources requested for launch at /home/nvidia/Downloads/pytorch/aten/src/THCUNN/generic/SpatialUpSamplingBilinear.cu:66
nvidia@tegra-ubuntu:~/Downloads/deep-learning-pytorch/segmentation$

Do you know what it could be?

解决方法是将pytorch源码中的 CUDA_NUM_THREADS =256