-
您好!我尝试使用您的CC聚类方法去替换传统deepcluster中K-means,我主体上使用的是您的代码,只是在cluster.py中去增加了一些部分,增加的部分是使用deepcluster的方式去训练。同时我使用了Market数据集,我之前已经在这个Market数据集上以CC的算法对模型进行预训练了(1000epoch),代码如下:
P.S.:被#框起来的部分是我新加进去的。
但是这段代码…
-
I'm trying to replace the cpu index by a gpu one but can't seem to do it on a distributed context.
Faiss version:
faiss 1.8.0 pypi_0 pypi
faiss-gpu …
-
-
Hi,sir,glad to see your job about the implementation of deepcluster by Facebook Research.
I also try to realize deepcluster but have some troubles with it.
What I am interest in is the implement of …
-
您好!没别的意思,with the respect, 只是我看到您论文里的实验部分的结果似乎比较一般,尤其是对一些类比较多的数据集。比如在tiny-imagenet,acc只能达到0.14,而据我所知meta在2018年搞出的deepcluster在imagenet已经可以做到0.35以上的acc了。所以就想问一下您这套深度聚类的方法优势究竟是什么呢?
P.S.:我算是个小白,只是单纯的有这个疑…
-
## 一言でいうと
End-to-Endで、CNNを利用したクラスタリングを行う研究。CNNで特徴抽出=>k-meansでクラスタリング=>そのラベルで学習、というステップを繰り返すことでクラスタリングを行う。ただ、無策で行うと巨大な1クラスタに全画像を収めて精度100%になる可能性があるため、空クラスタを防止したりラベルの偏りがないように工夫したりしている。
![image](htt…
-
Context: I have a dataset with say 1.000 classes and I want to perform K-means with gpu over each class.
E.g. (sketch):
``` Python
class KMeansModule:
def __init__(self, nb_classes, di…
-
When i run mask_generator.py ,it need some pre-train weights like /users/gyungin/sos/networks/pretrained/moco_v2_800ep_pretrain.pth.tar
Is it necessary? If so, how can i get it?
Thanks for your re…
-
Hello. I want to know whether this code implements the method of avoiding empty clusters in 2018ECCV. If so, can you point it out for me because I haven't seen this part.
Thanks
-
I download imagenet data through Kaggle and loaded the swav pretrained weights using the torch.hub.load method.
After that, for evaluation, we tried to evaluate the imagenet val set as a swav mo…