Telugu-POS-Tagger-and-Chunker

• This repository contains both Telugu pos tagger and chunker.
• Both the models are CRF based (Please install CRF++ toolkit. All the necessary details can be accesed from the below website: https://taku910.github.io/crfpp/).
• Consists of 2 steps.
  • In the first step, extract features from the raw sentences and run the POS tagger.
  • In the second step, the chunker model is run on the output of the POS tagger, here it only depends on the pos tags, not tokens.
• How to run the models:
  • sh run_telugu_pos_and_chunk_models_and_save_to_file.sh input_file pos_model chunk_model output_file
  • You can use the sample file in the repositoty to get the output: sh run_telugu_pos_and_chunk_models_and_save_to_file.sh telugu-input.txt telugu-pos-model.m telugu-chunk-model.m telugu-output.txt
• The raw sentences need to be tokenized, for Indian language tokenization use our tokenization repository (https://github.com/Pruthwik/Tokenizer_for_Indian_Languages)