This repository presents the final project of EECS6322 for Learning a Self-Expressive Networl for Subspace Clustering
Run the scripts by using following command
# dataset: MNIST, EMNIST, FashionMNIST, CIFAR10
python SENet_main.py --dataset=MNISTHere is some parameters that can be changed to get experiment results.
parser.add_argument('--batch_size', type=int, default=100)
parser.add_argument('--batch_eval', type=int, default=10000)
parser.add_argument('--dataset', type=str, default="CIFAR10")
parser.add_argument('--gamma', type=float, default=200.0)
parser.add_argument('--hid_dims', type=int, default=[1024, 1024, 1024])
parser.add_argument('--lmbd', type=float, default=0.9)
parser.add_argument('--lr', type=float, default=1e-3)
parser.add_argument('--lr_min', type=float, default=0.0)
parser.add_argument('--mean_subtract', dest='mean_subtraction', action='store_true')
parser.set_defaults(mean_subtraction=False)
parser.add_argument('--num_subspaces', type=int, default=10)
parser.add_argument('--out_dims', type=int, default=1024)
parser.add_argument('--spectral_dim', type=int, default=15)
parser.add_argument('--seed', type=int, default=0)
parser.add_argument('--total_iters', type=int, default=100000)
parser.add_argument('--top_k', type=int, default=1000)The processed experimental datasets can be download here. Put the datasets folder in the same work path to execute the code.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent