早上前2堂: RNN

X_Data = ['good', 'bad', 'worse', 'so good']
y = [1.0, 0.0, 0.0, 1.0]

count = 0
Encoding = dict()
for words in X_Data:
    for char in words:
        if char not in Encoding.keys():
            Encoding[char] = count+1
            count = count+1

X_squence = []
for words in X_Data:
    temp = []
    for char in words:
        temp.append(Encoding[char])
    X_squence.append(temp)

# 補數字將長度改成一樣
from tensorflow.keras.preprocessing.sequence import pad_sequences

X = pad_sequences(X_squence, maxlen=8, padding='post', value=0)

import numpy as np
y = np.array(y)

# 建構網路
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Embedding, SimpleRNN, Dense

RNN_Model = Sequential()
RNN_Model.add(Embedding(input_dim=11, output_dim=11, input_length=8))
RNN_Model.add(SimpleRNN(30))
RNN_Model.add(Dense(2))

# 訓練模型
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.losses import SparseCategoricalCrossentropy

adam_op = Adam(learning_rate=0.001)
loss_op = SparseCategoricalCrossentropy()

RNN_Model.compile(optimizer=adam_op, loss=loss_op, metrics=['acc'])

RNN_Model.fit(X, y, epochs=30, batch_size=2)

# 預測
np.argmax(RNN_Model.predict(X), axis=-1)

y

早上後2堂:

講解一些演算法，ex:爬山演算法，運用程式讓同學了解運作原理。

下午結業式 + 就業媒合