Introduction

This Repository was built to show case DistilBERT Implementation for NER which is built on Pytorch. We used distilbert-base-uncased Pretrained Model from Huggingface Transformer Library and Fine Tune it for token classification i.e. NER.

Since Guthub blocks, push for more than 100 MB's file (https://help.github.com/en/github/managing-large-files/conditions-for-large-files) therefore please download the Weights file from Google Drive ( https://drive.google.com/file/d/1MTdgl1qfOEo-TuT39ByrbQsFZQ-y7v3a/view?usp=sharing )

Output

Run Locally

Clone the project

  git clone https://github.com/nabarunbaruaAIML/BERT_NER.git

Go to the project directory

  cd BERT_NER

Install dependencies

  conda env create -f environment.yml

All dependencies are included in the environment.yml file.

Start the server

  Python3 clientApp.py