Learning rate
Overshooting : Learning rate의 값이 너무 커 튕겨나감
Takes too long : Learning rate의 값이 너무 작아 바닥까지 못내려감
Solution!
- Data(x) preprocessing
Original data -> Zero-centered-data -> Normalized data
- Standardiztion
X_std[i,0] = (X[i,0] - X[i,0].mean()) / X[i,0].std()
Overfitting
- Our model is very good with training data set
+ BUT not good at test dataset or in real use
Solution!
- More training data!
- Reduce the number of features
- Regularization
Regularization?
- Let's not have too big numbers in the weight
12reg = 0.001 * tf.reduce_sum(tf.square(w))
=================================================================
=================================================================
'Programming > Tensorflow' 카테고리의 다른 글
[Tensorflow] Binary Classfication (0) | 2017.07.30 |
---|---|
[Tensorflow] Multi-Variable linear regression (use matrix) (0) | 2017.07.30 |
[Tensorflow] Linear Regression (0) | 2017.07.30 |
[Tensorflow] 준비사항 2 (0) | 2017.07.30 |
[Tensorflow] 준비사항 1 (0) | 2017.07.30 |