Faceswap Forum

You may be able to train the Lightweight model on that card. It is specifically written for 2GB GPU cards, but it will be dependent on what else is using the GPU. Try it with a very low batchsize (2) and if it trains, you can try upping the batchsize in increments. If it doesn't train, try memory s...

My 2 cents... I had a similar lower end graphics and I stayed with Original Model. By adjusting coverage % to > 82% and opting ssim in loss function for training, i was able to get very good result. good for you. how many iterations and how long does it takes? i think i will look for a configuratio...

Unfortunately an 820m with 2gb of vram is unlikely to be viable. I'm afraid that you're going to have to rely on CPU or cloud training. ah, i see. thank you for the reply. are there any characteristics between viable gpu and non viable GPU? maybe you have a kind of list? oh yeah, about cloud traini...

hello, at first i thought it was okay to extract on CPU (because i cant make it run on GPU), but when i start to train, i realized it takes long time to compute. i've tried many solutions like: reinstall and tick on GPU (Nvidia) conda install tensorflow-gpu==1.15.0 install cuda cudnn from nvidia ins...