Multilinguals_1394

Sriharshitha Bondugula (2018111013)

Samartha S M (2018101094)

Mentor - Prashant Kodali

There are different jupyter notebook files in the src directory for different models.

To open any jupyter notebook file,

$ cd src
$ jupyter notebook model<N>.ipynb

where, N is the model ID as mentioned below,

N = 1 : Baseline model

N = 2 : Improved model 1 (using unk tokens)

N = 3 : Improved model 2 (using language information)

N = 4 : Improved model 3 (using both language information and unk tokens)

Some variables will have to be set in the jupyter notebook file before running the cells

Dataset path: File path which contains the dataset (separate for train and test set)

Model path: File path where you want to store the model instance

Outputs are in the folder /outputs.

Outputs of corresponding models are in /outputs/model_id

File that was used to create datasets is in /utils/create_datasets.py

Dataset link: https://iiitaphyd-my.sharepoint.com/:f:/g/personal/prashant_kodali_research_iiit_ac_in/EsSRXSdygL5DmyTaV-2oWxcB9t7NHA9oB5_ugYwp_HftYA?e=2iaTE1

Model checkpoints link: https://drive.google.com/file/d/187annbuwDM-N2xpMztZ0rtiOJXQMDP1w/view?usp=sharing