ML 2019 Fall Final Prj
~Use real images to predict drawing~ Arvin Liu
Real Image -> Drawing
(32 x 32 x 3)
(28 x 28 x 1)
Domain Adaptaion
有Label
沒有Label
Feature Extractor
Classifier
Source
Target
Source: blue
Target: Red
Feature Distribution
np.load("trainX.npy")# Get (5000, 32, 32, 3)
np.load("test.npy")# Get (100000, 28, 28, 1)
np.load("trainY.npy")# Get (5000,)
(label = 0~9)
(Balanced)
id,label0,01,02,03,04,05,06,0
就是同你的Hw3
Only torchvision's pre-trained models are allowed.
CLICK HERE
如果狀況太差的話會調降,只會降: )
Edge
Detection
直接讓Train很像Test也是種"Adaptation"?
*直接用cv2套件就可以了 (cv2.Canny)
這邊就有一個提示了XD
下圖是上圖的邊緣偵測結果
By Arvin Liu
ML2019 Final - Domain Adaptation 題目敘述