Python Cuda运行时内存不足错误-Pytorch使用8GB

我正在尝试用我的GTX 1080Ti训练一款具有11GB内存的机型，但Pytorch似乎为自己保留了约8GB内存，我可以通过任何方式减少这个数字吗？（这是最新的）

运行时错误：CUDA内存不足。尝试分配2.53 GiB（GPU 0；10.91 GiB总容量；4.15 GiB已分配；1.73 GiB空闲；PyTorch总共保留8.19 GiB）（malloc at/PyTorch/c10/cuda/CUDACacheGallocator.cpp:289）
帧#0:c10:：Error:：Error（c10:：SourceLocation，std:：string const&）+0x46（0x7f54646f7536 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libc10.so）
帧1:+0x1cf1e（0x7f5464940f1e in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libc10_cuda.so）
帧2:+0x1df9e（0x7f5464941f9e在/home/yasin/.local/lib/python3.6/site packages/torch/lib/libc10_cuda.so中）
帧#3:at:：native:：empty_cuda（c10:：ArrayRef，c10:：TensorOptions const&，c10:：可选）+0x135（0x7f54005f7535 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧4:+0xf7a66b（0x7f53febef66b in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧#5:+0xfc3f57（0x7f53fec38f57 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧6:+0x1075389（0x7f543b173389 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cpu.so）
帧#7:+0x10756c7（0x7f543b1736c7 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cpu.so）
第8帧：+0xf1557e（0x7f53feb8a57e in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧9:+0xf18043（0x7f53feb8d043 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cuda.so）
帧#10:at:：native:：cudnn_卷积(向后)输入（c10:：ArrayRef，at:：Tensor const&，at:：Tensor const&，c10:：ArrayRef，c10:：ArrayRef，long，bool，bool）+0xb2（0x7f53feb8d882 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cuda.so）
第11帧：+0xf7f3a0（0x7f53febf43a0 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧12:+0xfc3c38（0x7f53fec38c38 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cuda.so）
帧#13:at:：native:：cudnn_卷积(at:：Tensor const&，at:：Tensor const&，at:：Tensor const&，c10:：ArrayRef，c10:：ArrayRef，c10:：ArrayRef，long，bool，bool，std:：array）+0x4fa（0x7f53feb8ef1a in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cuda.so）
帧#14：+0xf7f6cb（0x7f53febf46cb in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧15:+0xfc3c94（0x7f53fec38c94 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cuda.so）
帧#16:+0x2c809b6（0x7f543cd7e9b6 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch_cpu.so）
帧17:+0x2cd0444（0x7f543cdce444 in/home/yasin/.local/lib/python3.6/site packages/torch/lib/libtorch\u cpu.so）
帧#18:torch:：autograd:：generated:：cudnconvolutionbackward:：apply（std:：vector）

你累了吗
torch.cuda.empty\u cache（）
？或者重新启动会话并重新运行代码？您使用的是哪种cuda和pytorch版本？我在cuda 10.2和pytorch 1.5上也遇到了同样的问题。解决这个问题是否顺利？在Windows中也存在这个问题，但讽刺的是，如果我使用较低规格的GPU在Google Colab中进行操作，它会起作用。这里的版本也是1.5。您累了吗
torch.cuda.empty_缓存（）
？或者重新启动您的会话并重新运行您的代码？您使用的是什么cuda和pytorch版本？我在cuda 10.2和pytorch 1.5上遇到了相同的问题。解决这个问题是否顺利？在Windows中也有这个问题，但讽刺的是，如果我在Google Colab中使用较低规格的GPU进行操作，它会起作用。这里也有版本1.5。
 RuntimeError: CUDA out of memory. Tried to allocate 2.53 GiB (GPU 0; 10.91 GiB total capacity; 4.15 GiB already allocated; 1.73 GiB free; 8.19 GiB reserved in total by PyTorch) (malloc at /pytorch/c10/cuda/CUDACachingAllocator.cpp:289)

frame #0: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x46 (0x7f54646f7536 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libc10.so)
    frame #1: <unknown function> + 0x1cf1e (0x7f5464940f1e in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libc10_cuda.so)
    frame #2: <unknown function> + 0x1df9e (0x7f5464941f9e in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libc10_cuda.so)
    frame #3: at::native::empty_cuda(c10::ArrayRef<long>, c10::TensorOptions const&, c10::optional<c10::MemoryFormat>) + 0x135 (0x7f54005f7535 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #4: <unknown function> + 0xf7a66b (0x7f53febef66b in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #5: <unknown function> + 0xfc3f57 (0x7f53fec38f57 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #6: <unknown function> + 0x1075389 (0x7f543b173389 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #7: <unknown function> + 0x10756c7 (0x7f543b1736c7 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #8: <unknown function> + 0xf1557e (0x7f53feb8a57e in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #9: <unknown function> + 0xf18043 (0x7f53feb8d043 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #10: at::native::cudnn_convolution_backward_input(c10::ArrayRef<long>, at::Tensor const&, at::Tensor const&, c10::ArrayRef<long>, c10::ArrayRef<long>, c10::ArrayRef<long>, long, bool, bool) + 0xb2 (0x7f53feb8d882 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #11: <unknown function> + 0xf7f3a0 (0x7f53febf43a0 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #12: <unknown function> + 0xfc3c38 (0x7f53fec38c38 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #13: at::native::cudnn_convolution_backward(at::Tensor const&, at::Tensor const&, at::Tensor const&, c10::ArrayRef<long>, c10::ArrayRef<long>, c10::ArrayRef<long>, long, bool, bool, std::array<bool, 2ul>) + 0x4fa (0x7f53feb8ef1a in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #14: <unknown function> + 0xf7f6cb (0x7f53febf46cb in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #15: <unknown function> + 0xfc3c94 (0x7f53fec38c94 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cuda.so)
    frame #16: <unknown function> + 0x2c809b6 (0x7f543cd7e9b6 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #17: <unknown function> + 0x2cd0444 (0x7f543cdce444 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #18: torch::autograd::generated::CudnnConvolutionBackward::apply(std::vector<at::Tensor, std::allocator<at::Tensor> >&&) + 0x378 (0x7f543c996918 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #19: <unknown function> + 0x2d89c05 (0x7f543ce87c05 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #20: torch::autograd::Engine::evaluate_function(std::shared_ptr<torch::autograd::GraphTask>&, torch::autograd::Node*, torch::autograd::InputBuffer&) + 0x16f3 (0x7f543ce84f03 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #21: torch::autograd::Engine::thread_main(std::shared_ptr<torch::autograd::GraphTask> const&, bool) + 0x3d2 (0x7f543ce85ce2 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #22: torch::autograd::Engine::thread_init(int) + 0x39 (0x7f543ce7e359 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_cpu.so)
    frame #23: torch::autograd::python::PythonEngine::thread_init(int) + 0x38 (0x7f545f97b4d8 in /home/yasin/.local/lib/python3.6/site-packages/torch/lib/libtorch_python.so)
    frame #24: <unknown function> + 0xbd6df (0x7f54675c86df in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
    frame #25: <unknown function> + 0x76db (0x7f546c2ff6db in /lib/x86_64-linux-gnu/libpthread.so.0)
    frame #26: clone + 0x3f (0x7f546c63888f in /lib/x86_64-linux-gnu/libc.so.6)