Part of Speech Tagging for the Persian dataset using ParsBert and Multilingual BERT
The repository contains Jupyter notebooks for testing different aspects of a POS-Tagging BERT model. The notebooks include:
-
Test 1 - ParsBert.ipynb: This notebook tests the ParsBert model on the given dataset.
-
Test 2 - Increase Learning Rate - bad.ipynb: This notebook tests the effect of increasing the learning rate on the performance of the model.
-
Test 3 - Decrease Weight decay.ipynb: This notebook tests the effect of decreasing the weight decay on the performance of the model.
-
Test 4 - MultiLingual.ipynb: This notebook tests the multilingual Bert model on the dataset instead of using ParsBert.
-
Test 5 - MultiLingual uncased.ipynb: This notebook tests the uncased multilingual Bert model on the dataset.
-
Test 6 - ParsBert v2.ipynb: This notebook tests the ParsBert v2 model on the given dataset.
In addition to the notebooks, the repository also contains a README.md file and a Report.pdf file. The README.md file provides information about the repository, including a brief overview of the model and instructions on how to run the notebooks. The Report.pdf file provides a more detailed analysis of the model's performance and the explanation of this exercise.
You can access the dataset in the Dataset folder.
Overall, the repository provides a good starting point for anyone interested in using a POS-Tagging BERT model. The notebooks are well-organized and easy to follow, and the report provides valuable insights into the model's performance.
Just so you know, you should adjust the used paths because the code has been written on Google Colab.