# pomelo-single
###### tags: `紀錄`

categories 6
tain 2603
test 530

## fig explanation
b Training accurarcy
g Validation accurarcy
r Training loss
y Validation loss

## Effectient B0 <mark>看起來over fitting</mark>
batch size 32
epoch 90
Loss 3.396
Acc 0.564
:::warning
validation_split=0.2 -> 會over fitting
所以改為 x_train, x_valid, y_train, y_valid = train_test_split(x_train , y_train , test_size=0.2, shuffle= True)
:::

---
## EfficientNet B0 <mark>ing</mark>

Loss 0.285
Accc 0.930
## DenseNet 121

Loss 0.645
Accc 0.781