very interested in more PyTorch lessons! Your Keras TF video series is wonderful. If someone like you could bring that depth and scope of lessons for PyTorch I would be so happy :)
Yes, Sir. I like to see more on PyTorch. and sir have ever see intel's effort on open source tensor compiler called PlaidML? it would be great, if can you comment on it.
I am having this error while running the installation command for cpu ERROR: Could not find a version that satisfies the requirement torch==1.8.1+cpu ERROR: No matching distribution found for torch==1.8.1+cpu please help me with this
return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse) RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0! (when checking argument for argument index in method wrapper__index_select) resolution ???? xformers , medvram or lowvram ???? and speed?
i have rtx 3070 when i exucute (nvidia smi) i get cuda version 12 how can i get pytorch 12 version and wich cuda version should i install and cudann and toolkit plz
i have Geforce GTX 1650 (VRAM= 4 GB) and i have some problem with installation cuda toolkit for tensorflow. i had install Tf = 2.4, and cuda toolkit = 11.0 , cuDNN= 8.0. But it is not working . So i decided to learn with pytorch.
@@HeatonResearch I am having the same issue I am using a miniconda, a python 3.9.1, CUDA 11.0 toolkit, I did not find any pytorch compatible with CUDA 11.0 on pytorch.org, there is a version for CUDA 11.1, I wanted to retain my CUDA 11.0 and download pytorch is it feasible and if yes can you please tell me how?
I'm getting close: Tensor Flow Version: 2.3.0 Keras Version: 2.4.0 Python 3.8.3 (default, May 19 2020, 06:50:17) [MSC v.1916 64 bit (AMD64)] Pandas 1.1.0 Scikit-Learn 0.23.2 GPU is NOT AVAILABLE (But I installed the Nvidia driver last, error)
I have really enjoyed keeping the class as Tensorflow, since I’ve been following for a few years now... but it seems to be the way of the world now. You should do a video converting a model between Keras and pyTorch.
Thanks for the great video. Please do more in Pytorch, especially reinforcement learning. I have watched your tubetube on migrating of TF1 to TF2. Please do a short video on convert TF1 to TF2 with using tf_upgrade_v2. I have been trying it for a few months, but still can't get it right. Thanks again.
So do you want to see more PyTorch from me? If so, like this video and let me know in the comments! That is how I know.
Don't I have to install a specific verison of CUDA & Cudnn like I did with Tensorflow?
Yes, please. I've been curious about PyTorch especially with NLU using Rasa. Thanks again for sharing.
Sure thing!Yes, I've been looking at Rasa a bit.
Hi there. Thanks for creating this video. It has helped me set up both tensorflow and now pytorch on my pc.
That was very helpful and easy to follow. Thank you Jeff.
very interested in more PyTorch lessons! Your Keras TF video series is wonderful. If someone like you could bring that depth and scope of lessons for PyTorch I would be so happy :)
Yes, Sir. I like to see more on PyTorch.
and sir have ever see intel's effort on open source tensor compiler called PlaidML? it would be great, if can you comment on it.
I have not seen PlaidML, I will take a look, though.
Getting more pytorch course is always great! looking forward to it
Great to hear! Thanks.
More PyTorch content will be awesome.
Well, that was easy! Where's all the TensorFlow-like problems identifying the card and libraries?
Prof, i am in (torch) env and paste conda env update --file tools.yml . file not found error. any suggestion? TQ
I'm not sure >cd t81_558_deep_learning.master ?
I am having this error while running the installation command for cpu
ERROR: Could not find a version that satisfies the requirement torch==1.8.1+cpu
ERROR: No matching distribution found for torch==1.8.1+cpu
please help me with this
You have provided very useful content. Thank you sir.
Dose GPU cuda confined in torch envs so it will not have effect tensorflow-gpu I installed earlier?
Yes, ideally. That is why I like to use conda ENV's rather than system wide drivers.
hi sir. in the last step the error "ModuleNotFoundError: No module named 'torch' come out. how can i solve?
Thank you sir for this amazing video
return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0! (when checking argument for argument index in method wrapper__index_select)
resolution ????
xformers , medvram or lowvram ???? and speed?
i have rtx 3070 when i exucute (nvidia smi) i get cuda version 12 how can i get pytorch 12 version and wich cuda version should i install and cudann and toolkit plz
So I don't need to pre install CUDA and cudnn, conda will take care of it all?
The only thing I need is nvidia drivers?
Will this currently work with notebook, are there any changes to the code?
Hi
I have AMD 6900XT and wanted to utilise it, how can I set pytorch directml for it
Thanks
Hi Jeff, thanks for your helpful videos. today(2 october) with version windows 20226 the gpu over wsl2 does not work. Do you know some fix?. thanks
I have NVIDIA, the output GPU not available.
Your title is to install Pytorch, and your document is to install tensorflow, confused!
Hey, the link to github is for tensorflow, not pytorch. Thats confusing. But we love you!
My laptop is using AMD radeon, is it possible for me to install PyTorch?
thank you very much
can you recommend good CUDA programming books :)
Hello every one, why torch dont work for spyder ??
Are there any different in installing jupyter lab?
No, for Jupyter it is the same as TF.
Yes Sir
i have Geforce GTX 1650 (VRAM= 4 GB) and i have some problem with installation cuda toolkit for tensorflow. i had install Tf = 2.4, and cuda toolkit = 11.0 , cuDNN= 8.0. But it is not working . So i decided to learn with pytorch.
PyTorch is much easier to install than TensorFlow.
@@HeatonResearch I am having the same issue I am using a miniconda, a python 3.9.1, CUDA 11.0 toolkit, I did not find any pytorch compatible with CUDA 11.0 on pytorch.org, there is a version for CUDA 11.1, I wanted to retain my CUDA 11.0 and download pytorch is it feasible and if yes can you please tell me how?
I'm getting close:
Tensor Flow Version: 2.3.0
Keras Version: 2.4.0
Python 3.8.3 (default, May 19 2020, 06:50:17) [MSC v.1916 64 bit (AMD64)]
Pandas 1.1.0
Scikit-Learn 0.23.2
GPU is NOT AVAILABLE
(But I installed the Nvidia driver last, error)
The NVIDIA driver is the most painful part. Mine quit working after my last Windows update a few days ago, I need to figure out what happened.
@@HeatonResearch You should consider switching from Windows to Linux. The Microsoft upadte policy is a bad joke.
the points to tensorflow installation but not pytorch
Sir can I install pytorch without private conda environment on windows?
Yes, you can install it to the root environment if you want. Less control that way, but it works.
@@HeatonResearch Thank you Sir
I have really enjoyed keeping the class as Tensorflow, since I’ve been following for a few years now... but it seems to be the way of the world now. You should do a video converting a model between Keras and pyTorch.
Thanks for the great video.
Please do more in Pytorch, especially reinforcement learning.
I have watched your tubetube on migrating of TF1 to TF2. Please do a short video on convert TF1 to TF2 with using tf_upgrade_v2. I have been trying it for a few months, but still can't get it right.
Thanks again.
Thanks for the suggestions. Reinforcement learning in particular is where I think PyTorch is moving ahead.