Update README.md

README.md

@@ -13,20 +13,6 @@ This repository contains code for a Nepali Sentiment Analysis model. The model i
 
 ## Dataset
 
-The dataset used for training this model is sourced from [Shushant/NepaliSentiment](https://huggingface.co/datasets/Shushant/NepaliSentiment) on the Hugging Face Datasets Hub. The dataset consists of labeled Nepali text samples with corresponding sentiment labels.
-
-## Model Architecture
-
-The model architecture used in this repository is based on the BERT (Bidirectional Encoder Representations from Transformers) model, specifically the `bert-base-multilingual-cased` variant. BERT is a powerful pre-trained language model that can be fine-tuned for various natural language processing tasks, including sentiment analysis.
-
-The model takes Nepali text as input and produces sentiment predictions as output. It uses tokenization techniques to convert the text into numerical representations and leverages the transformer-based architecture to capture contextual relationships between words in the text.
-
-# Nepali Sentiment Analysis
-
-This repository contains code for a Nepali Sentiment Analysis model. The model is trained to predict the sentiment (positive, negative, or neutral) of Nepali text.
-
-## Dataset
-
 The dataset used for training this model is sourced from [Shushant/NepaliSentiment](https://huggingface.co/datasets/Shushant/NepaliSentiment) on the Hugging Face Datasets Hub. The dataset consists of labelled Nepali text samples with corresponding sentiment labels.
 
 ## Model Architecture
@@ -49,10 +35,20 @@ To use the Nepali Sentiment Analysis model:
 
 5. Evaluate the model's performance using suitable metrics such as accuracy, precision, recall, or F1-score.
 
+6. To try the model practically, open `nepali-sentiment-model-tryout.ipynb` in Jupyter Notebook.
+
+7. In the `model-tryout.ipynb` file, provide the string you want to analyze for sentiment.
+
+8. Run the cells in `model-tryout.ipynb` to obtain the predicted sentiment label for the provided string.
+
+9. Optionally, you can also modify the code to analyze multiple strings or perform a batch analysis.
+
 ## Repository Structure
 
-- `model_training.ipynb`: Jupyter Notebook containing the code for training the Nepali Sentiment Analysis model.
+- `model-training.ipynb`: Jupyter Notebook containing the code for training the Nepali Sentiment Analysis model.
+- `model-tryout.ipnyb` : Hupyter Notebook containing the code to tryout the Model
 - `inference.ipynb`: Jupyter Notebook demonstrating how to perform inference using the trained model.
+- `nepali-sentiment-model-tryout.ipynb`: Jupyter Notebook for trying the model practically with a string input.
 - `requirements.txt`: List of required dependencies for running the code.
 - `README.md`: This readme file provides an overview of the repository.
 
@@ -74,7 +70,6 @@ During the training process, the model achieved the following accuracies for dif
 ![image](https://github.com/Sibindra/nepali-sentiment-analysis-model/assets/59206903/19bc55b3-4fab-4e73-a535-b5c6712f3029)
 
 
-
 To improve the accuracy of the model, you can consider making the following changes:
 
 - Adjust the batch size: Try experimenting with different batch sizes to see if it has any impact on the model's performance. Sometimes, a smaller or larger batch size can lead to better results.
@@ -87,4 +82,4 @@ I encourage open-source contributors to explore ways to improve the accuracy of
 
 ## License
 
-This project is licensed under the [MIT License](LICENSE).
+This project is licensed under the [MIT License](LICENSE).