update some extensions, and when I restarted stable. prune.global_unstructured when I use prune.global_unstructure I get that error please help What is the purpose of this D-shaped ring at the base of the tongue on my hiking boots? """, def __init__(self, num_classes, pretrained=False): super(C3D, self).__init__() self.conv1 = nn.quantized.Conv3d(3, 64, kernel_size=(3, 3, 3), padding=(1, 1, 1))#..54.14ms self.pool1 = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2)), self.conv2 = nn.quantized.Conv3d(64, 128, kernel_size=(3, 3, 3), padding=(1, 1, 1))#**395.749ms** self.pool2 = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)), self.conv3a = nn.quantized.Conv3d(128, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1))#..208.237ms self.conv3b = nn.quantized.Conv3d(256, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1))#***..348.491ms*** self.pool3 = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)), self.conv4a = nn.quantized.Conv3d(256, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1))#..64.714ms self.conv4b = nn.quantized.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1))#..169.855ms self.pool4 = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)), self.conv5a = nn.quantized.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1))#.27.173ms self.conv5b = nn.quantized.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1))#.25.972ms self.pool5 = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2), padding=(0, 1, 1)), self.fc6 = nn.Linear(8192, 4096)#21.852ms self.fc7 = nn.Linear(4096, 4096)#.10.288ms self.fc8 = nn.Linear(4096, num_classes)#0.023ms, self.relu = nn.ReLU() self.softmax = nn.Softmax(dim=1), x = self.relu(self.conv1(x)) x = least_squares(self.pool1(x)), x = self.relu(self.conv2(x)) x = least_squares(self.pool2(x)), x = self.relu(self.conv3a(x)) x = self.relu(self.conv3b(x)) x = least_squares(self.pool3(x)), x = self.relu(self.conv4a(x)) x = self.relu(self.conv4b(x)) x = least_squares(self.pool4(x)), x = self.relu(self.conv5a(x)) x = self.relu(self.conv5b(x)) x = least_squares(self.pool5(x)), x = x.view(-1, 8192) x = self.relu(self.fc6(x)) x = self.dropout(x) x = self.relu(self.fc7(x)) x = self.dropout(x), def __init_weight(self): for m in self.modules(): if isinstance(m, nn.Conv3d): init.xavier_normal_(m.weight.data) init.constant_(m.bias.data, 0.01) elif isinstance(m, nn.Linear): init.xavier_normal_(m.weight.data) init.constant_(m.bias.data, 0.01), import torch.nn.utils.prune as prunedevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")model = C3D(num_classes=2).to(device=device)prune.random_unstructured(module, name="weight", amount=0.3), parameters_to_prune = ( (model.conv2, 'weight'), (model.conv3a, 'weight'), (model.conv3b, 'weight'), (model.conv4a, 'weight'), (model.conv4b, 'weight'), (model.conv5a, 'weight'), (model.conv5b, 'weight'), (model.fc6, 'weight'), (model.fc7, 'weight'), (model.fc8, 'weight'),), prune.global_unstructured( parameters_to_prune, pruning_method=prune.L1Unstructured, amount=0.2), --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) in 19 parameters_to_prune, 20 pruning_method=prune.L1Unstructured, ---> 21 amount=0.2 22 ) ~/.local/lib/python3.7/site-packages/torch/nn/utils/prune.py in global_unstructured(parameters, pruning_method, **kwargs) 1017 1018 # flatten parameter values to consider them all at once in global pruning -> 1019 t = torch.nn.utils.parameters_to_vector([getattr(*p) for p in parameters]) 1020 # similarly, flatten the masks (if they exist), or use a flattened vector 1021 # of 1s of the same dimensions as t ~/.local/lib/python3.7/site-packages/torch/nn/utils/convert_parameters.py in parameters_to_vector(parameters) 18 for param in parameters: 19 # Ensure the parameters are located in the same device ---> 20 param_device = _check_param_device(param, param_device) 21 22 vec.append(param.view(-1)) ~/.local/lib/python3.7/site-packages/torch/nn/utils/convert_parameters.py in _check_param_device(param, old_param_device) 71 # Meet the first parameter 72 if old_param_device is None: ---> 73 old_param_device = param.get_device() if param.is_cuda else -1 74 else: 75 warn = False AttributeError: 'function' object has no attribute 'is_cuda', prune.global_unstructured when I use prune.global_unstructure I get that error. I got this error when working with Pytorch 1.12, but the error eliminated with Pytorch 1.10. No, 1.13 is out, thanks for confirming @kurtamohler. AttributeError: module 'torch' has no attribute 'is_cuda' How to handle a hobby that makes income in US, Linear Algebra - Linear transformation question. Hi Franck, Thanks for the update. To figure out the exact issue we need your code and steps to test from our end.Could you share the entire code an We tried running your code.The issue seems to be with the quantized.Conv3d, instead you can use normal convolution3d. Will Gnome 43 be included in the upgrades of 22.04 Jammy? To figure out the exact issue we need yourcode and steps to test from our end.Could you sharethe entire code and steps in a zip file? The error is unfortunately not super descriptive or guiding me how to fix it. Do roots of these polynomials approach the negative of the Euler-Mascheroni constant? Well occasionally send you account related emails. It seems part of these problems have been solved and the data is automatically downloaded when I run the codes. My code is GPL licensed, can I issue a license to have my code be distributed in a specific MIT licensed project? - the incident has nothing to do with me; can I use this this way? Can I tell police to wait and call a lawyer when served with a search warrant? Site design / logo 2023 Stack Exchange Inc; user contributions licensed under CC BY-SA. Making statements based on opinion; back them up with references or personal experience. Why does it seem like I am losing IP addresses after subnetting with the subnet mask of 255.255.255.192/26? GCC version: (Ubuntu 11.3.0-1ubuntu1~22.04) 11.3.0 . Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug 1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)] Module I'm running without dreambooth now as I had to use CPU training anyway with my 4Gb card and they made that harder recently so I'd gone to Colab, which is much quicker anyway. [pip3] numpy==1.23.4 To learn more, see our tips on writing great answers. AttributeError: module 'torch' has no attribute 'cuda', update some extensions, and when I restarted stable. Follow Up: struct sockaddr storage initialization by network format-string, Full text of the 'Sri Mahalakshmi Dhyanam & Stotram'. By clicking Accept all cookies, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy. You just need to find the Thank you. Find centralized, trusted content and collaborate around the technologies you use most. Asking for help, clarification, or responding to other answers. to your account. You may try updating. The default one installed is only with CPU support. Browse other questions tagged, Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide. Did any DOS compatibility layers exist for any UNIX-like systems before DOS started to become outmoded? Steps to reproduce the problem. See instructions here https://pytorch.org/get-started/locally/ The nature of simulating nature: A Q&A with IBM Quantum researcher Dr. Jamie We've added a "Necessary cookies only" option to the cookie consent popup. So probably you either have somewhere used torch.float in your code or you have imported some code with torch.float. Not the answer you're looking for? module if update to an extension did this, please let us know - in my book, that kind of behavior is borderline hostile as an extension should NOT change core libraries, only libraries that are extra for that extension. By clicking Post Your Answer, you agree to our terms of service, privacy policy and cookie policy. rev2023.3.3.43278. please downgrade (or upgrade) to the latest version of 3.10 Python stderr: Traceback (most recent call last): But I meet the following problems and it seems difficult for me to fix it by myself: the main error is "AttributeError: module 'torch._C' has no attribute '_cuda_setDevice'". To subscribe to this RSS feed, copy and paste this URL into your RSS reader. ), Implement Seek on /dev/stdin file descriptor in Rust. Implement Seek on /dev/stdin file descriptor in Rust. With the more extensive dataset, I receive the AttributeError in the subject header and RuntimeError: Pin memory threat exited unexpectedly after 8 iterations. By clicking Sign up for GitHub, you agree to our terms of service and Have you installed the CUDA version of pytorch? profile. update some extensions, and when I restarted stable. torch.cuda.amptorch1.6torch1.4 1.7.1 You may re-send via your, AttributeError: module 'torch' has no attribute 'is_cuda', Intel Connectivity Research Program (Private), oneAPI Registration, Download, Licensing and Installation, Intel Trusted Execution Technology (Intel TXT), Intel QuickAssist Technology (Intel QAT), Gaming on Intel Processors with Intel Graphics. You may re-send via your If thats not possible, and assuming you are using the GPU, use torch.cuda.amp.autocast. The nature of simulating nature: A Q&A with IBM Quantum researcher Dr. Jamie We've added a "Necessary cookies only" option to the cookie consent popup. module Why do we calculate the second half of frequencies in DFT? torch cannot detect cuda anymore, most likely you'll need to reinstall torch. Try removing it then reinstalling. Why are Suriname, Belize, and Guinea-Bissau classified as "Small Island Developing States"? "After the incident", I started to be more careful not to trip over things. I have same error after install pytorch from channel "soumith" with this command: After reinstalling from pytorch channel all works fine. For more complete information about compiler optimizations, see our Optimization Notice. Yesterday I installed Pytorch with "conda install pytorch torchvision -c pytorch". 'numpy.ndarray' object has no attribute 'cuda' - PyTorch Forums Site design / logo 2023 Stack Exchange Inc; user contributions licensed under CC BY-SA. For the code you've posted it makes no sense. Hi, Thank you for posting your questions. First of all use torch.cuda.is_available() to detemine the CUDA availability also we need more details . Sign up for a free GitHub account to open an issue and contact its maintainers and the community. torch.cuda.amp is available in the nightly binaries, so you would have to update. vegan) just to try it, does this inconvenience the caterers and staff? . To figure out the exact issue we need yourcode and steps to test from our end.Could you sharethe entire code and steps in a zip file? Have a question about this project? You may just comment it out. Normal boot up. cuDNN version: Could not collect I could fix this on the 1.12 branch, but will there be a 1.12.2 release? Error: " 'dict' object has no attribute 'iteritems' ", Getting Nan result out of ResNet101 backbone with Kitti images. Connect and share knowledge within a single location that is structured and easy to search. Is there a single-word adjective for "having exceptionally strong moral principles"? You have to call the decorator as given in the docs and examples: Powered by Discourse, best viewed with JavaScript enabled, Older version of PyTorch: with torch.autocast('cuda'): AttributeError: module 'torch' has no attribute 'autocast'. Pytorch Simple Linear Sigmoid Network not learning. Later in the night i did the same and got the same error. Seemed to resolve it for the other people on that thread earlier too. WebLKML Archive on lore.kernel.org help / color / mirror / Atom feed * [PATCH v38 00/39] LSM: Module stacking for AppArmor [not found] <[email protected]> @ 2022-09-27 19:53 ` Casey Schaufler 2022-09-27 19:53 ` [PATCH v38 01/39] LSM: Identify modules by more than name Casey Schaufler ` (38 more replies) 0 siblings, The name of the source file was 'torch.py'. Traceback (most recent call last): We are closing the case assuming that your issue got resolved.Please raise a new thread in case of any further issues. ERROR: Could not find a version that satisfies the requirement torch==1.13.1+cu117 (from versions: none) Thanks! CMake version: version 3.22.1 RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available () is Fal. pytorch1.61.6 please help I just sent the iynb model So something is definitely hostile as you said =P. File "C:\ai\stable-diffusion-webui\launch.py", line 105, in run How can I import a module dynamically given the full path? CUDA Already on GitHub? I have two machines that I need to check my code across one is Ubuntu 18.04 and the other is Ubuntu 20.04. How do I check if an object has an attribute? Also happened to me and dreambooth was one of the ones that updated! How can I import a module dynamically given the full path? If you don't want to update or if you are not able to do so for some reason. Try to transform the numpy array to a tensor before calling tensor.cuda () It should install the latest version. Webimport torch.nn.utils.prune as prune device = torch.device("cuda" if torch.cuda.is_available() else "cpu") model = C3D(num_classes=2).to(device=device) RuntimeError: Error running command. run_python("import torch; assert torch.cuda.is_available(), 'Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check'") How can this new ban on drag possibly be considered constitutional? I don't think the function torch._C._cuda_setDevice or torch.cuda.set_device is available in a cpu-only build. Using Kolmogorov complexity to measure difficulty of problems? Is there a single-word adjective for "having exceptionally strong moral principles"? AttributeError: module 'torch' has no attribute 'is_cuda' rev2023.3.3.43278. Re:AttributeError: module 'torch' has no attribute AttributeError: module 'torch' has no attribute 'is_cuda', Intel Connectivity Research Program (Private), oneAPI Registration, Download, Licensing and Installation, Intel Trusted Execution Technology (Intel TXT), Intel QuickAssist Technology (Intel QAT), Gaming on Intel Processors with Intel Graphics. At this moment we are not planning to move to pytorch 1.13 yet. Help for those needing help starting or connecting to the Intel DevCloud, The Intel sign-in experience has changed to support enhanced security controls. In torch.distributed, how to average gradients on different GPUs correctly? In following the Pytorch tutorial at https://pytorch.org/tutorials/beginner/deep_learning_60min_blitz.html. To subscribe to this RSS feed, copy and paste this URL into your RSS reader. --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) in 1 get_ipython().system('pip3 install torch==1.2.0+cu92 torchvision==0.4.0+cu92 -f https://download.pytorch.org/whl/torch_stable.html') ----> 2 torch.is_cuda AttributeError: module 'torch' has no attribute 'is_cuda'. Libc version: glibc-2.35, Python version: 3.8.15 (default, Oct 12 2022, 19:15:16) [GCC 11.2.0] (64-bit runtime) . Asking for help, clarification, or responding to other answers. Can you provide the full error stack trace? The nature of simulating nature: A Q&A with IBM Quantum researcher Dr. Jamie We've added a "Necessary cookies only" option to the cookie consent popup. CUDA used to build PyTorch: 11.6 Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide. Why is this sentence from The Great Gatsby grammatical? Why does Mister Mxyzptlk need to have a weakness in the comics? Just renamed it to something else and delete the file named 'torch.py' in the directory The text was updated successfully, but these errors were encountered: torch cannot detect cuda anymore, most likely you'll need to reinstall torch. profile. [pip3] torch==1.12.1+cu116 AttributeError: module torch.cuda has no attribute amp By clicking Accept all cookies, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy. MIOpen runtime version: N/A Command: "C:\ai\stable-diffusion-webui\venv\Scripts\python.exe" -c "import torch; assert torch.cuda.is_available(), 'Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check'" I'm using Windows, conda environment, installed Pytorch-1.7.1, Torchvision-0.8.2, Cuda-Toolkit-11.0 > all compatible. Pytorchpthh5python AttributeError: 'module' object has no attribute 'dumps'Keras Thanks for your answer. with torch.autocast ('cuda'): AttributeError: module 'torch' has no attribute 'autocast' I have this version of PyTorch on Ubuntu 20.04: python Python 3.8.10 (default, Is there a single-word adjective for "having exceptionally strong moral principles"? If you don't want to update or if you are not able to do so for some reason. AttributeError: module 'torch._C' has no attribute '_cuda_setDevice' facebookresearch/detr#346 marco-rudolph mentioned this issue on Sep 1, 2021 error https://pytorch.org/. I had to delete my venv folder in the end and let automatic1111 rebuild it. WebAttributeError: module 'torch' has no attribute 'cuda' Press any key to continue . Please click the verification link in your email. module 'torch.cuda' has no attribute '_UntypedStorage'. Shouldn't it be pip unistall torch and pip install torch? . File "C:\ai\stable-diffusion-webui\launch.py", line 360, in Older version of PyTorch: with torch.autocast('cuda'): Asking for help, clarification, or responding to other answers. Still get this error--module 'torch._C' has no attribute '_cuda_setDevice', https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/360, https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/67, https://github.com/samet-akcay/ganomaly/blob/master/options.py#L40, module 'torch._C' has no attribute '_cuda_setDevice', AttributeError: module 'torch._C' has no attribute '_cuda_setDevice'. As the PyTorch forum member with the most posts manages the PyTorch Core team @ NVIDIA. didnt work as well. If you encounter an error with "RuntimeError: Couldn't install torch." Is XNNPACK available: True, Versions of relevant libraries: Sorry, you must verify to complete this action. to your account, On a machine with PyTorch version: 1.12.1+cu116, running the following code gets error message module 'torch.cuda' has no attribute '_UntypedStorage'. Traceback (most recent call last): Powered by Discourse, best viewed with JavaScript enabled, AttributeError: module 'torch.cuda' has no attribute '_UntypedStorage'. CUDA_MODULE_LOADING set to: GPU models and configuration: GPU 0: NVIDIA GeForce RTX 3090 What browsers do you use to PyTorch - "Attribute Error: module 'torch' has no attribute 'float', How Intuit democratizes AI development across teams through reusability. Sorry, you must verify to complete this action. File "C:\ai\stable-diffusion-webui\launch.py", line 129, in run_python microsoft/Bringing-Old-Photos-Back-to-Life#100. What's the difference between a Python module and a Python package? CUDA runtime version: Could not collect How do I check if an object has an attribute? Since this issue is not related to Intel Devcloud can we close the case? What else should I do to get right running? You signed in with another tab or window. [notice] A new release of pip available: 22.3 -> 23.0.1 To learn more, see our tips on writing great answers. yes I reported an issue yesterday and met with much the same response. I am actually pruning my model using a particular torch library for pruning, device = torch.device("cuda" if torch.cuda.is_available() else "cpu")class C3D(nn.Module): """ The C3D network. conda list torch gives me: But, when asking for the torchvision version in Jupyter Notebook, I get: AttributeError: module 'torch.fft' has no attribute 'fftfreq' Hot Network Questions Add circled letters in titles In your code example I cannot find anything like it. prepare_environment() Connect and share knowledge within a single location that is structured and easy to search. Please click the verification link in your email. How do I unload (reload) a Python module? [Bug]: AttributeError: module 'torch' has no attribute venv "C:\ai\stable-diffusion-webui\venv\Scripts\Python.exe" However, the error disappears if not using cuda. By clicking Accept all cookies, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy. WebAttributeError: module 'torch' has no attribute 'cuda' Press any key to continue . torch torch.rfft torch.irfft torch.rfft rfft ,torch.irfft irfft I'm stuck with this issue and the problem is I cannot use the latest version of pytorch (currently using 1.12+cu11.3). However, some new errors appear as follows: And I wonder that if it may be impossible to run these codes in the cpu only computer? The same code can run correctly on a different machine with PyTorch version: 1.8.2+cu111, Collecting environment information Staging Ground Beta 1 Recap, and Reviewers needed for Beta 2, Calling a function of a module by using its name (a string). So for example when changing in the imported code: torch.tensor([1, 0, 0, 0, 1, 0], dtype=torch.float) to torch.FloatTensor([1,0,0,0,1,0]) it might still complain about torch.float even if the line then doesn't contain a torch.floatanymore (it even shows the new code in the traceback). I just got the following error when attempting to use amp. Site design / logo 2023 Stack Exchange Inc; user contributions licensed under CC BY-SA. Why does Mister Mxyzptlk need to have a weakness in the comics? The cuda () method is defined for tensors, while it seems you are calling it on a numpy array. rev2023.3.3.43278. First of all usetorch.cuda.is_available() to detemine the CUDA availability also weneed more details tofigure out the issue.Could you provide us the commands and stepsyou followed? How do I check if an object has an attribute? So if there was an error in the old code this error might still occur and the traceback then points to the line you have just corrected. No issues running the same script for a different dataset. WebAttributeError: module tensorflow has no attribute GPUOptionsTensorflow 1.X 2.XTensorflow 1.Xgpu_options = tf.GPUOptions(per_process_gpu_memory_fraction)Tensorflow 2.Xgpu_options =tf.compat.v1.GPUOptions(per_process_gpu_memory_fractio Traceback (most recent call last): File "D:/anaconda/envs/ml/Lib/site-packages/torch_sparse/__init__.py", line 4, in import torch File "D:\anaconda\envs\ml\lib\site-packages\torch_, File "D:\anaconda\envs\ml\lib\platform.py", line 897, in system return uname().system File "D:\anaconda\envs\ml\lib\platform.py", line 785, in uname node = _node() File "D:\anaconda\envs\ml\lib\platform.py", line 588, in _node import socket File "D:\anaconda\envs\ml\lib\socket.py", line 52, in import os, sys, io, selectors, File "D:\anaconda\envs\ml\lib\selectors.py", line 12, in import select File "D:\anaconda\envs\ml\Lib\site-packages\torch_sparse\select.py", line 1, in from torch_sparse.tensor import SparseTensor File "D:\anaconda\envs\ml\lib\site-packages\torch_sparse_. Making statements based on opinion; back them up with references or personal experience. or can I please get some context of why this is occuring? I tried to fix this problems by refering https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/360 and https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/67 Command: "C:\ai\stable-diffusion-webui\venv\Scripts\python.exe" -m pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 --extra-index-url https://download.pytorch.org/whl/cu117 i actually reported that to dreambooth extension author 3 weeks ago and got told off. We are closing the case assuming that your issue got resolved.Please raise a new thread in case of any further issues. Yes twice updates to dreambooth have screwed my python environment badly. class GradScaler(torch.cuda.amp.GradScaler): AttributeError: module torch.cuda has no attribute amp Environment: GPU : RTX 8000 CUDA: 10.0 Pytorch How can we prove that the supernatural or paranormal doesn't exist? AttributeError: 'datetime' module has no attribute 'strptime', Error: " 'dict' object has no attribute 'iteritems' ". Please see. Whats the grammar of "For those whose stories they are"? AttributeError:partially initialized module 'torch' has no attribute 'cuda' Ask Question Asked Viewed 894 times 0 In the __init__.py of the module named torch [pip3] torchaudio==0.12.1+cu116 Have a question about this project? I tried to reproduce the code from https://github.com/samet-akcay/ganomaly and run the commands in the git bash software. It seems that you need to add --device cpu in the command line to make it work. ROCM used to build PyTorch: N/A, OS: Ubuntu 22.04.1 LTS (x86_64) To subscribe to this RSS feed, copy and paste this URL into your RSS reader. By clicking Sign up for GitHub, you agree to our terms of service and Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/cu117 Have a question about this project? NVIDIA doesnt develop, maintain, or support pytorch. Can we reopen this issue and maybe get a backport to 1.12? If you have a line like in the example you've linked, it makes perfectly sense to get an error like this. Similarly to the line you posted in your question. Steps to reproduce the problem. In my case command looks like: But you must obtain package list for yours machine form this site: Does ZnSO4 + H2 at high pressure reverses to Zn + H2SO4? [pip3] torchvision==0.13.1+cu116 Commit where the problem happens. vegan) just to try it, does this inconvenience the caterers and staff? Clang version: Could not collect Browse other questions tagged, Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide. For the Nozomi from Shinagawa to Osaka, say on a Saturday afternoon, would tickets/seats typically be available - or would you need to book? Why do I get AttributeError: 'NoneType' object has no attribute 'something'? Sorry for late response that is, I change the code torch.cuda.set_device(self.opt.gpu_ids[0]) to torch.cuda.set_device(self.opt.gpu_ids[-1]) and torch._C._cuda_setDevice(device) to torch._C._cuda_setDevice(-1)but it still not works.