anuragreddygv323 / Bahasa-NLP-Tensorflow

Gathers Tensorflow deep learning models for Bahasa Malaysia NLP problems

---

Bahasa-NLP-Tensorflow, Gathers Tensorflow deep learning models for Bahasa Malaysia NLP problems, code simplify inside Jupyter Notebooks 100% including dataset.

Table of contents

• Augmentation
• Sparse classification
• Long-text classification
• Dependency Parsing
• Entity Tagging
• Abstractive Summarization
• Extractive Summarization
• POS Tagging
• Optical Character Recognition
• Question-Answer
• Speech to Text
• Stemming
• Topic Generator
• Text to Speech
• Topic Modeling
• Word Vector

Augmentation

1. word2vec Malaya

Sparse classification

1. Fast-text Ngrams

Normal-text classification

1. Fast-text
2. Only Attention

70+ more models can get from here.

Long-text classification

1. Dilated CNN
2. Wavenet

Dependency Parsing

1. Bidirectional LSTM + CRF
2. Bidirectional LSTM + CRF + Bahdanau
3. Bidirectional LSTM + CRF + Luong

Entity Tagging

1. Bidirectional LSTM + CRF
2. Bidirectional LSTM + CRF + Bahdanau
3. Bidirectional LSTM + CRF + Luong

POS Tagging

1. Bidirectional LSTM + CRF
2. Bidirectional LSTM + CRF + Bahdanau
3. Bidirectional LSTM + CRF + Luong

Abstractive Summarization

1. Dilated Seq2Seq
2. Pointer Generator + Bahdanau Attention
3. Pointer Generator + Luong Attention

Extractive Summarization

1. Skip-thought
2. Residual Network + Bahdanau Attention

Optical Character Recognition

1. CNN + LSTM RNN

Question-Answer

1. End-to-End + GRU
2. Dynamic Memory + GRU

Speech to Text

1. BiRNN + LSTM + CTC Greedy
2. Wavenet
3. Deep speech 2

Text to Speech

1. Tacotron
2. Seq2Seq + Bahdanau Attention
3. Deep CNN + Monothonic Attention + Dilated CNN vocoder

Stemming

1. Seq2seq + Beam decoder
2. Seq2seq + Bahdanau Attention + Beam decoder
3. Seq2seq + Luong Attention + Beam decoder

Topic Generator

1. TAT-LSTM
2. TAV-LSTM
3. MTA-LSTM

Topic Modeling

1. Lda2Vec

Word Vector

1. word2vec
2. ELMO
3. Fast-text