I have a 1650 GPU graphics card and have installed all the drivers and necessary components, but CUDA is not working. What do I need to do? “1650 with max-q design”
Same experience; have the GTX 1650 and CUDA installed but not working! Was hoping to get some pointers if this is possible, but no conclusive answers so it seems
1 Like
Same for me.
torch.cuda.is_available()
False
GeForce GTX1650 (it lists 1024 CUDA cores)
NVCUDA64.dll v31.0.15.3114
1 Like
any solution to this yet?
is any one figured this out?
Hey everyone,
I am a fresher data scientist. I was trying to do model training of Yolov8m model on my system, that has a GTX 1650. Cuda 12.1 was installed with pytorch and its showing when I do the version check, but still while training the model it is not supporting and the loss values are ‘nan’ and map values are 0. I even tried installing the cuda toolkit 12.3 from Nvidia, still no good.
Can someone please guide me what the problem is and how I can solve it.
Thank you.
I confirm GPU is working with CUDA for GeForce 1650. (tested with TensorFlow following the a bit outdated “Native Windows” TensorFlow installation (but didn’t succeed with WSL on newer version of TensorFlow))
Nvidia staff should update its list.
I just installed the desktop version of the GTX 1650. My goal is to use it for the FastAI course, pytorch, and general cuda+python coding to learn. Can anyone who got it to work tell me if there is a specific version I need to download?
I am using this version from the nvidia site: cuda_12.4.0_551.61_windows.exe
1 Like
i have seen here: here
that GTX 1650 is compatible with CUDA 7.5
GeForce GTX 1650 is CUDA-capable. Here is what I did on my windows 11. Installed Visual Studio build tools 2022 (for me v17.8.3) and updated GPU drivers. (For me Studio version 552.22). Then checked the supported CUDA versions for PyTorch, and installed it(CUDA 12.1). Finally, installed PyTorch for CUDA 12.1 from here.
which OS are you using? I am using windows 11, will it work on that?
where did you check the supported cuda version? i am facing the same issue. but my nvidia driver is 12.6 and cuda is also 12.6.
Did you set any paths?
My OS is also Windows 11. Your nvidia drivers might be unsupported with higher CUDA 12.6. First uninstall Cuda 12.6. Then type
nvidia-smi
in cmd for the driver version and check the supported CUDA version. Then download and install it.
1 Like
did you use libtorch or another version of pytorch? I am trying to build libtorch on visual studio and gpu never gets used.
how can i fix it.
Go through the steps I mentioned. Once drivers are updated, I would suggest you CUDA 12.1. Once you install it, follow the steps properly.