AI說書 - 從0開始 - 288 | Tokenizer 重要性範例之 Embedding 訓練

我想要一天分享一點「LLM從底層堆疊的技術」，並且每篇文章長度控制在三分鐘以內，讓大家不會壓力太大，但是又能夠每天成長一點。

sample = open("text.txt", "r")
s = sample.read()

f = s.replace("\n", " ")
data = []

for i in sent_tokenize(f):
	temp = []
	for j in word_tokenize(i):
		temp.append(j.lower())
	data.append(temp)
	
	
# Creating Skip Gram model
model2 = gensim.models.Word2Vec(data, min_count = 1, vector_size = 512, window = 5, sg = 1)
print(model2)

window = 5 限制輸入句子中當前單字和預測單字之間的距離，結果為：

為了要檢視效果好壞，我們撰寫一隻計算 Cosine 相似度的程式：

def similarity(word1, word2):
	cosine = False
	try:
		a = model2[word1]
		cosine = True
	except KeyError:
		print(word1, ":[unk] key not found in dictionary")
	
	try:
		b = model2[word2]
	except KeyError:
		cosine = False
		print(word2, ":[unk] key not found in dictionary")
	
	if(cosine == True):
		dot = np.dot(a, b)
		norma = np.linalg.norm(a)
		normb = np.linalg.norm(b)
		cos = dot / (norma * normb)
		
		aa = a.reshape(1,512)
		ba = b.reshape(1,512)
		cos_lib = cosine_similarity(aa, ba)
	if(cosine == False):
		cos_lib = 0
	return cos_lib