Blade6570 / Learningimage-to-imagetranslationusingpairedandunpairedtrainingsamples

Learning image-to-image translation using paired and unpaired training samples
https://tutvision.github.io/Learning-image-to-image-translation-using-paired-and-unpaired-training-samples/
Other
19 stars 3 forks source link

getting training error Please help #5

Open shivom9713 opened 3 years ago

shivom9713 commented 3 years ago

error

I have added the dataset in the mentioned format ... but i got this error Please Help

Blade6570 commented 3 years ago

Hi, which PyTorch version are you using? This code is written for V0.3.1.

shivom9713 commented 3 years ago

I used google colab....i dont actually remember the version ...but i will try using the version you mentioned once again Thanks

On Mon, Nov 30, 2020, 2:18 PM Soumya Tripathy notifications@github.com wrote:

Hi, which PyTorch version are you using? This code is tested on V0.3.1.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Blade6570/Learningimage-to-imagetranslationusingpairedandunpairedtrainingsamples/issues/5#issuecomment-735645220, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANNEXP27RZEALETMBUARSI3SSNL45ANCNFSM4UFZVFLQ .

shivom9713 commented 3 years ago

torch>=1.4.0 torchvision>=0.5.0 dominate>=2.4.0 visdom>=0.1.8.8

Sir i have used the above requirements and tried torch 1.1.0 and torchvision 0.3.0 separately I got an error 23

Please help me to get the code work

Blade6570 commented 3 years ago

Hi, Seems like some package version mismatch problem. Please set up your enviornemnt again by keeping Python at 3.5.4 Pytorch at 0.3.1 PIL at 6.1 and let me know if the problem persists. You are using pytorch 1.1.0 which is not compatible for this code base.

shivom9713 commented 3 years ago

caca

Still shows error even after installing the required versions of pillow and torch sir my email id is shivom9713@gmail.com I am doing my Mtech project....Your code can be helpful for my project Please mail me so that i can contact you directly Thanks

shivom9713 commented 3 years ago

semi This time i ran the code with : torch 0.3.1 torchvision 0.2 dominate and visdom

some GPU error is thrown by the complier

Blade6570 commented 3 years ago

I think you have installed pytorch compiled with cuda-8 but in your google colab they have cuda-9. So you need a torch v0.3 with cuda-9 support. Something like https://download.pytorch.org/whl/cu90/torch-0.3.0.post2-cp35-cp35m-linux_x86_64.whl. You can download this file and then install by pip install [downloaded file]

shivom9713 commented 3 years ago

Sir, i tried a lot... Can we schedule a zoom/Google Meet if you are available tomorrow or any other day . It will be a great help for me. I am working on Cycle GAN for infrared image to visible image translation and i was thinking to try combining PIX2PIX and Cycle GAN objective and I found your paper to be relevant . A guidance / direction from your side will be very beneficial for me in my project work . Thanks and Regards, SHIVOM BHARGAVA MTECH IIT Kanpur

On Tue, Jan 5, 2021 at 5:35 PM Soumya Tripathy notifications@github.com wrote:

I think you have installed pytorch compiled with cuda-8 but in your google colab they have cuda-9. So you need a torch v0.3 with cuda-9 support. Something like https://download.pytorch.org/whl/cu90/torch-0.3.0.post2-cp35-cp35m-linux_x86_64.whl. You can download this file and then install by pip install [downloaded file]

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Blade6570/Learningimage-to-imagetranslationusingpairedandunpairedtrainingsamples/issues/5#issuecomment-754594685, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANNEXPY53HVBPV25OGLK5G3SYL52HANCNFSM4UFZVFLQ .

kiyaaabdian commented 2 years ago

if you don't want answer the question, why do you write For any help, please contact us at ?

Blade6570 commented 2 years ago

@kiyaaabdian What is the issue btw? I don't work in university anymore so have limited access to university emails. If it is version mismatch problem like the issue of @shivom9713 then difficult for me to replicate.