Hard to put dataset in the right place for just a test.

Vandermode / ERRNet

Single Image Reflection Removal Exploiting Misaligned Training Data and Network Enhancements (CVPR 2019)

MIT License

247 stars 52 forks source link

Hard to put dataset in the right place for just a test. #8

Open aragorn1025 opened 5 years ago

aragorn1025 commented 5 years ago

Is there any hint for put dataset in the right place for at least a test?

MenSanYan commented 5 years ago

@Vandermode @aragorn1025 What should I do for testing only one image? Can you provide an example for inference ? Anticipate for your reply.

97k commented 5 years ago

Hi @MenSanYan, I was also trying to test for one single image. I did the following to get the results.

Created a folder, placed my testing image inside that directory
Commented all code related to the creation of dataset and data loader
Created my dataset for example, test_ref_dataset = datasets.RealDataset('/home/aditya/project/ERRNet/test_ref/')
Created a dataloader for the same
test_ref_dataloader = datasets.DataLoader( test_ref_dataset, batch_size=8, shuffle=False, num_workers=opt.nThreads, pin_memory=True)
res = engine.test(test_ref_dataloader, savedir=result_dir) After this, I was able to get for a single image in the results directory.

MohammadSamaraChowagiken commented 3 years ago

Can we use this network for commercial purpose for a small project, as a part of an image processing processes.

innat commented 3 years ago

To inference on raw samples. In test_errnet.py, right after line 17, commented rest of the part. And put below code

engine = Engine(opt)
test_ref_dataset = datasets.RealDataset('/content/my_img/') # where you save your image

test_ref_dataloader = datasets.DataLoader(
                    test_ref_dataset, batch_size=8, shuffle=False,
                    num_workers=opt.nThreads, pin_memory=True)

res = engine.test(test_ref_dataloader, savedir="/content/") # where result would be saved

WGHao commented 2 years ago

Can i use the code without ubuntu?