adobe/NLP-Cube

Name: NLP-Cube

Owner: Adobe Systems Incorporated

Description: Natural Language Processing Pipeline - Sentence Splitting, Tokenization, Lemmatization, Part-of-speech Tagging and Dependency Parsing

Created: 2018-03-13 18:50:44.0

Updated: 2018-05-21 18:25:01.0

Pushed: 2018-05-21 13:59:50.0

Homepage:

Size: 6606

Language: Python

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

NLP-Cube

Setup:

Before running the server, you need the model's weights, and you can follow two approaches to get them:

• Download data in order to train the model yourself
• Download already existing model weights

Installing dyNET:

1. Make sure you have Mercurial, python, pip, cmake installed (you can also check steps documented here)

2. Install Intel's MKL library

3. Install dyNET by using the installation steps from the manual installation page. More specifically, you should use:

   install cython
   r dynet-base
   ynet-base
   
   clone https://github.com/clab/dynet.git
   lone https://bitbucket.org/eigen/eigen -r 2355b22  # -r NUM specified a known working revision
   
   ynet
   r build
   uild
   e .. -DEIGEN3_INCLUDE_DIR=/path/to/eigen -DMKL_ROOT=/opt/intel/mkl -DPYTHON=`which python2`
   
    -j 2 # replace 2 with the number of available cores
    install
   
   ython
   on2 ../../setup.py build --build-dir=.. --skip-build install
   

Training the lemmatizer:

Use the following command to train your lemmatizer:

 Running the server:

the following command to run the server locally:

Current status

• we treat words and character embeddings in a similar fashion
• we tested with character encodings only (feature cutoff is set at 100)

ToDO

• [ ] provide training examples
• [x] add word embeddings
• [x] find a good network achitecture for POS tagging
• [x] prepare a neural/based language pipeline
• [ ] pre-train models using universal dependencies
• [x] add a parser

Parser architecture

-----------------                    -------------------------- 
|word emebddings|----          ------|morphological embeddings|
-----------------    |        |      --------------------------
                     |        |
                   --------------
                   |concatenate |
                   --------------
                          |
                  ----------------
                  |bdlstm_1_layer|
                  ----------------
                          |
                  ----------------                  
                  |bdlstm_2_layer| 
                  ----------------                    
                          |-----------------------------------------------------------------                          
                  ----------------                                                         |
                  |bdlstm_3_layer|                                                         |
                  ----------------                                                         |
                          |                                                                |
     ---------------------------------------------                    ---------------------------------------------              
     |           |                |              |                    |           |                |              |
     |           |                |              |                    |           |                |              |
 ---------  -----------       ----------    ------------          ---------  -----------       ----------    ------------
 |to_link|  |from_link|       |to_label|    |from_label|          |to_link|  |from_link|       |to_label|    |from_label|
 ---------  -----------       ----------    ------------          ---------  -----------       ----------    ------------
      |        |                      |       |                       |           |                  |            |
    --------------                 ---------------                  ------------------            -------------------
    |softmax link|                 |softmax label|                  |aux softmax link|            |aux softmax label|
    --------------                 ---------------                  ------------------            -------------------

Tagger architecture

-----------------                    ---------------------- 
|word emebddings|----          ------|character embeddings|
-----------------    |        |      ----------------------
                     |        |
                   --------------
                   |tanh_1_layer|
                   --------------
                          |
                  ----------------
                  |bdlstm_1_layer|
                  ----------------
                          |
                   --------------                  
                   |tanh_2_layer|-------------------
                   --------------                   |
                          |                         |
                  ----------------         -------------------
                  |bdlstm_2_layer|         |aux_softmax_layer|
                  ----------------         -------------------
                          |
                   ---------------
                   |softmax_layer|
                   ---------------