# baseline


# 1.將過去上次EWC

1.74.8
2.75.69
3.75.79

# 2. 去除EWC(更新teacher model
1.76.43

2.77.09

# 3. 每次都將teacher 設為上一次的student model

# 4. 更新 base model，測試EWC base的效果如何
1.

2.
3.
4.
# 4.不將train_x越變越小
1.
2.
3.
4.

# 5.lambda 值測試
## 0.001
1.
2.
3.
4.

### 0.0001

# mixup

## alpha = 0.2
## 0.5
1.
2.
3.
4.

5.
## 5
1.
2.
3.
4.
5.

## 1
1.