DeepEnergyModels

Word Segmentation in Sanskrit Using Energy Based Models

Reviewer Copy

We are in the process of organising the program files.

Getting Started

Please download the 2 compressed files 'dir.zip' and 'wordsegmentation.rar' to your working directory and extract them into folders named 'dir' and 'wordsegmentation' respectively.

Your working directory should be as follows

• Working Directory
  • wordsegmentation
    • skt_dcs_DS.bz2_4K_bigram_mir_10K
    • skt_dcs_DS.bz2_4K_bigram_mir_heldout
  • dir

Prerequisites

• Python3
  • scipy
  • numpy
  • csv
  • pickle
  • multiprocessing
  • bz2

Instructions for Training

Change your current directory to 'dir'

Run the file Train_clique.py by using the following command

• python Train_clique.py

To train on different input features like BM2,BM3,BR2,BR3,PM2,PM3,PR,PR3 please modify the bz2_input_folder value in the main function before beginning the training.

Instructions for Testing

After training, please modify the 'modelList' dictionary in 'test_clique.py' with the name of the neural network that has been saved during training. While testing for a feature, please provide the name of the neural net which was trained for the same feature.

We only provide the trained model for the feature BM2 which was our best performing feature. If the name of the neural net is not changed, then the testing will be performed on the pre-trained model for BM2 provided in outputs/train_t7978754709018

To test with a particular feature vector use the tag of the feature while execution

• python test_clique.py -t

For example:

• python test_clique.py -t BM2

After finishing the testing please run the following command to see the precision and recall values for both the word and word++ prediction tasks

• python evaluate.py

For example:

• python evaluate.py BM2