Comments (14)
Have you checked the dimension of CCH?
from contrastive-clustering.
I have set thd dimension of CCH with class_num
from contrastive-clustering.
After 20 epochs, i tested 'ImageNet-dogs',i print len(indices) and i get 1,but 'CIFAR10/100' is True
from contrastive-clustering.
from contrastive-clustering.
Perhaps it is because the file organization of ImageNet-dogs is wrong. You may check the ground-truth labels and see if it contains 0-14.
from contrastive-clustering.
I get the datasets from http://vision.stanford.edu/aditya86/ImageNetDogs/,it includes images and annotations .
How do i deal with?
Thanks.
from contrastive-clustering.
I command using ImageFolder following the instructions in https://pytorch.org/vision/stable/datasets.html#imagefolder.
from contrastive-clustering.
The dataset you downloaded seems to be inconsistent with the one used in the paper. In our experiments, the ImageNet-dogs data consists of 15 subclasses of the original Imagenet dataset, the indices of which are given in /datasets/ImageNet-dogs.txt.
from contrastive-clustering.
I have already corrected the data consists of 15subclasses, and then I need to merge 'images' with 'annotations'?
from contrastive-clustering.
Please follow the instructions in https://pytorch.org/vision/stable/datasets.html#imagefolder and arrange images of different classes in different subfolders.
from contrastive-clustering.
Now,i followed the indices and arrange images of different classes in in different subfolders with the indices,but i only get 2436 images...
from contrastive-clustering.
Each class contains 1300 images and thus there should be 19500 images in total.
from contrastive-clustering.
The datasets with 120 classes downloaded from http://vision.stanford.edu/aditya86/ImageNetDogs, now, I process this dataset according to the index, but per class only has more than 100.
from contrastive-clustering.
Perhaps you need to download the ImageNet dataset and select its 15 subclasses.
from contrastive-clustering.
Related Issues (20)
- Visualization HOT 3
- Low Accuracy and NMI HOT 3
- 关于baseline的疑问 HOT 2
- 关于代码的运行 HOT 8
- cifar10和cifar00 HOT 2
- loss 大小 HOT 8
- RuntimeError: numel: integer multiplication overflow HOT 8
- 关于idea的Fig1行级别解释 HOT 2
- 关于防止平凡解的计算
- 关于STL-10 HOT 2
- resnet模型参数的疑问 HOT 2
- customized dataset HOT 1
- Test metrics don't match HOT 9
- Can I ditch feature extraction and use the method HOT 2
- 专用GPU内存占满,batchsize达不到256 HOT 4
- 您好,我将您的代码改了一下用于预测路网交通流量,取一小时的时间序列进行节点间比较,但是结果是把所有的节点划分成一类,想问一下您有什么建议吗 HOT 14
- 关于x_i和x_j是怎么得到的 HOT 2
- 当数据量较少时可以获得不错的聚类效果,但是当数据量增大时训练到100轮之后聚类结果则会逐渐下降 HOT 2
- 推理时候使用k-means HOT 1
- pytorch version and torchvision version are not compatib HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from contrastive-clustering.