Home.md

 # Torchvision_sunner
-### The flexible extension of torchvision
+### The flexible extension of torchvision toward multiple image space
 
 [![Packagist](https://img.shields.io/badge/Version-18.4.4-yellow.svg)]()
 [![Packagist](https://img.shields.io/badge/Pytorch-0.3.0-red.svg)]()
@@ -30,47 +30,7 @@ Import library first
 import torchvision_sunner.transforms as sunnertransforms
 import torchvision_sunner.data as sunnerData
 ```
-Form dataset:
-```python
-# Load single image folder
-dataset = sunnerData.ImageDataset(
-    root_list = ['./waiting_for_you_dataset/real_world'],
-)
-
-# Load multiple image folder
-dataset = sunnerData.ImageDataset(
-    root_list = ['./waiting_for_you_dataset/real_world', './waiting_for_you_dataset/wait'],
-)
-```
-
-Extra transpose method
-```python
-dataset = sunnerData.ImageDataset(
-    root_list = ['./waiting_for_you_dataset/real_world', './waiting_for_you_dataset/wait'],
-    transform = transforms.Compose([
-        sunnertransforms.Rescale((160, 320)),
-        sunnertransforms.ToTensor(),
-
-        # BHWC -> BCHW
-        sunnertransforms.Transpose(sunnertransforms.BHWC2BCHW),
-        sunnertransforms.Normalize([127., 127., 127.], [127., 127., 127.])
-    ]) 
-)
-```
-
-The wrapper of usual `Dataloader`, but it can know the whole number of batch!
-```python
-loader = sunnerData.ImageLoader(dataset, batch_size=32, shuffle=False, num_workers = 2)
-```
-
-Transfer the tensor into numpy, and it's similar to the usage of `ImageDataset`!
-```python
-batch_img = sunnertransforms.tensor2Numpy(batch_img, transform = transforms.Compose([
-    sunnertransforms.UnNormalize([127., 127., 127.], [127., 127., 127.]),
-    sunnertransforms.Transpose(sunnertransforms.BCHW2BHWC),
-]))
-```
-
+And start for your flexible data processing! The more detail can be refer to [wiki](https://github.com/SunnerLi/Torchvision_sunner/wiki).     
 Notice
 ---
 * This package provides two backend image processing library working: opencv and skimage. Since the opencv can show the continuous image easily, the default library we use is opencv. On the contrary, the installation of opencv is tedious. You can choose skimage to become the backend library while it can be easily installed.