Комментарии:
Hi I am getting error torch is not compiled with cuda enabled in my jupyter notebook how to get rid of this
ОтветитьIs it mandatory to install cuDNN files??
ОтветитьPytorch is not compatible with few nvidia like 3080,3080ti and all. Could you please share the details which are nvidia gpu cards support both pytorch and tensorflow.
ОтветитьFew points I would like to add:
1) If your torch package is already installed, then first uninstall that torch package and then install the CUDA supported torch package else when you will try to install that CUDA torch package, the pip might refuse to install it (because you already have torch installed).
2) Even if your CUDA version and torch version are incompatible, still the CUDA-torch package may work. e.g: My CUDA version was 12.1 and the torch library I installed was compatible with CUDA 11.8 and still it worked well for CUDA 12.1.
can we keep two versions of cuda? I need cuda 10 for one project which already runs, and now i need to install pytorch 1.5+ for that i need to upgrade cuda but I want to keep cuda 10 as well. Moreover, when I run nvidia-smi, i get cuda 12 version and when running nvcc --version, i get cuda 10. Please help.
ОтветитьThank you SO much!! This is the only tutorial that helped me!!
Ответитьthx worked
Ответитьi have this err torch not compiled with cuda enabled
ОтветитьCan't we install pytorch in base environment??? is it mendatory to create a new env??
ОтветитьThank you so much, I have been trying to install pytorch and i had gone through so many video tutorials but only this video tutorial helped me to install it Sir. Great job
Ответитьwas getting annoyed with cuda before watching this video, a bundle of thanks
ОтветитьDo we need Visual studio 2019?
ОтветитьVs code me ye program kese run karr
(
Import torch
....
...
Thnx 👍
Awesome video ( subscribed)❣️
i dont have nvidia gpu..i have AMD Radeon HD520 gpu.. can i use cuda? and how? selecting computer platform to CPU instead of CUda???
ОтветитьI have an NVIDIA GeForce RTX 3050Ti graphic card in my laptop, the ouptut for the codes you given were showing like this:
```
>>> torch.cuda.is_available()
False
>>> torch.cuda.current_device()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\Users\csd\.conda\envs\cuda_pytorch\lib\site-packages\torch\cuda\__init__.py", line 481, in current_device
_lazy_init()
File "C:\Users\csd\.conda\envs\cuda_pytorch\lib\site-packages\torch\cuda\__init__.py", line 210, in _lazy_init
raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled
>>> torch.cuda.device_count()
0
```
Any idea what could be wrong ??
Sir mery pas nvidia ka GPU hai phir bhi cuda install Nahi horaha , cuda 9,10,11 koi b Nahi horaha...
Just error graphic driver not compatible, but driver bhi same install hain...
bhai mere pas dedicated NVidia ka GPU nahi hai , mere pas interl ka UDH 620 GPU hai , please batao pytorch kaise install kare ??
ОтветитьWhy need cuda
Ответить