## Week 4 * 之前都用模擬資料,這次運用真實資料(sunspot data) * Combine CNN & LSTM * 精進法1: 在不同epoch下觀察learning rate,看他在哪裡收斂 * 精進法2: 加多epochs次數,即使大圖看起來已平緩,拉近看還是有差 * 精進法3: 嘗試bidirectional LSTMs 雙向預測e.g. 從前後文互相推測(有些data不能用此邏輯) 此資料實驗結果overfitting,所以此資料還是要單向比較好 * 精進法4: 調整batch size 原本預設是32,後來嘗試各種值16最好 * 精進法5: 用Mini batch gradient descent優化(影片沒多說)
×
Sign in
Email
Password
Forgot password
or
By clicking below, you agree to our
terms of service
.
Sign in via Facebook
Sign in via Twitter
Sign in via GitHub
Sign in via Dropbox
Sign in with Wallet
Wallet (
)
Connect another wallet
New to HackMD?
Sign up