Roman
Hi! Thank you for your contribution! Could you please change the default values to `.png` so the default behaviour is not changed after your fix?
Hi! Sorry for the leate reply! You can try to turn off the validation - then the pipeline would use less gpu memory and load will be a bit more...
[This link](https://pytorch-lightning.readthedocs.io/en/latest/accelerators/gpu.html#horovod) might be also useful
Just set very long period of validation, by adjusting `.yaml` or by adding a command line argument to train.py, e.g. `trainer.kwargs.check_val_every_n_epoch=10000` However, when turning validation off you have to adjust...
Hi! Could you please point to the places we use `plt.imread`? I do no see any in our main code
Hi! All LaMa models in the paper were trained using 256x256 crops from Places. Original resolution of images in Places is approximately 512
Feel free to reopen the issue if you have further questions
Due to the nature of convolutions, the networks adapt to the scale of objects/textures - and perform best on exactly that scale. Inpainting in 256 is not very interesting in...
Hi! Thank you for the links! Hama looks great! I wonder, if your app has an interactive segmentation technology, which refines contours of the object to be deleted? Also, if...
@kjsik11 I've added a link to Hama to our readme! Thank you! Is my description of Hama fair?