semisupervised-FGSBIR
semisupervised-FGSBIR copied to clipboard
[CVPR 2021] "More Photos are All You Need: Semi-Supervised Learning for Fine-Grained Sketch Based Image Retrieval”, IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2021.
I'm sorry to bother you I would like to know what kind of environment your code runs in, such as the Python version, Pytorch version, etc
dataset
Hi, the number of sketch images is 6648/1275 from link https://drive.google.com/file/d/1frltfiEd9ymnODZFHYrbg741kfys1rq1/view?usp=sharing and https://drive.google.com/u/0/uc?id=15s2BR-QwLgX_DObQBrYlUlZqUU90EL9G&export=download, but in your paper is 6730/2000 of Shoev2 and Chairv2. Did i put down the wrong dataset?...
I wonder where the codes for the FG-SBIR task are? It seems like all the released codes are for the sketch generating task. How do you get the reword from...
Each picture has multiple sketches. When training sketch generator, how to form training sample pairs. Is each sketch will be used paired with corresponding picture? Or each picture with one...
Hello again, I have found that some of the requirements for reproduction are missing: 1) the pre-trained model used in `model.load_state_dict(torch.load('./modelCVPR21/QMUL/model_photo2Sketch_QMUL_2Dattention_8000_.pth'))` is not provided, 2) validation or testing environments are...