Follow me on M E D I U M: towardsdatascience.com/likeli...
STRATASCRATCH
For Thousands of Data Science Interview Questions + Solutions, sign up for stratascratch: www.stratascratch.com/?via=Co...
REFERENCES
[1] Why it's okay to add position embeddings: randorithms.com/2020/11/17/Ad...
[2] Main Transformer Paper: arxiv.org/abs/1706.03762
[3] Word2Vec Vs Transformers: www.quora.com/What-are-the-ma...
[4] Using sub-words in BERT: handsonnlpmodelreview.quora.c...
[5] In High Dimensinos, randomly drawn vectors are orthogonal: math.stackexchange.com/questi...
[6] Stackexchage answer on Positional encodings: datascience.stackexchange.com...
[7] Good information on positional encoding: kazemnejad.com/blog/transform...
[8] Master Positional Encoding: towardsdatascience.com/master...
[9] Reddit Thread on Positional Encoding: www.reddit.com/r/MachineLearn...
TIMESTAMPS
0:00 Introduction
0:44 Transformer Architecture
1:54 Data Science Interview Sponsor
3:04 Vectors
5:09 Role of Vectors in Transformers
7:07 Position Encoding
10:35 Multi Head Attention
11:37 Vector Operations: Addition Vs Concatenation
13:42 Beyond Transformers (BERT, Sentence Transformer)
Негізгі бет Vectors In Transformer Neural Networks
Пікірлер: 53