DIS
DIS copied to clipboard
xuebinqin
This is the repo for our new project Highly Accurate Dichotomous Image Segmentation
Hello ! Thank you for the paper, It's a great work :) But I wonder know how to enable DistributedDataParallel training? Whether to consider supporting this feature in the future...
Hello, thank you again for the great work: it made the processes a lot faster for us. I was wondering if there is a way of getting rid of staircases...
Hello ! Thank you for the paper, really liked it :) I'm wondering why you decided to monitor validation metrics while training DIS since you didn't do it for U2Net...
Thank you for the great repository! Is it possible to pass batches of images for inference? I noticed none of the examples showed batch inference, but I am wondering if...
Hey @xuebinqin! 👋 Thank you for uploading the code and trained networks. It works easily and is very usefull. This pull request makes it possible to run your model inside...
class ISNetDIS(nn.Module): In this class, the hx variable is not used, is it a code error?
I tried overfitting on fewer images [24] to see if issue is real, f1 error array and hence np.amax(f1) both are coming out to be 0. Any solutions?
error
/root/miniconda3/envs/ISNet-cu118/lib/python3.8/site-packages/torch/nn/_reduction.py:42: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead. warnings.warn(warning.format(ret)) building model... batch size: 8 --- create training dataloader --- ------------------------------ train -------------------------------- --->>> train dataset...
net = ISNetGTEncoder() #UNETGTENCODERCombine() 和 hypar["model"] = ISNetDIS() #U2NETFASTFEATURESUP()报红，但是不影响运行，当跑到训练时报torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 0 has a total capacity of 12.00 GiB of which 0...
