Telugu_movie_review_sentiment
This model is a fine-tuned version of xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3627
- Accuracy: 0.8814
- F1: 0.8889
Model description
This project is also related to Natural Language Processing (NLP). It is also related to the Telugu dataset, which is very useful for people who only know the Telugu language. In this project, we developed a Sentiment Analysis Language Model with a rating generator. We created a rating generator that will give a rating for the movie review according to the model's output probabilities. Based on the positive and negative polarity probability obtained from the movie review, a rating can be generated for the movie. For this project, my role is to create a dataset. We take one sentence from the movie and hand-label it with the words "positive" and "negative". During the training, the model will be able to understand which statements are positive and negative. and further work is being done on documents
Intended uses & limitations Intended Uses: Sentiment Analysis for Telugu Movie Reviews: This project is designed for Telugu-speaking users, allowing them to analyze the sentiment of Telugu movie reviews. Movie Rating Generator: It provides a quick movie rating based on sentiment analysis results, assisting Telugu-speaking audiences in decision-making. Training Data Creation: This project offers a manually labeled dataset for Telugu movie reviews, aiding NLP research in the field of sentiment analysis. Documentation: Comprehensive documentation is available to guide users and developers.
Limitations:
Language Restriction: Only supports the Telugu language. Sentiment Accuracy: Accuracy depends on the input text's quality and complexity. Dataset Bias: Accuracy is influenced by biases in the training data. Movie-Specific Content: This may not handle reviews with obscure references well. Model Updates: Requires regular updates to maintain accuracy. External Factors: Accuracy may be affected by user bias and industry changes.
Training procedure
Step-1: The corpus "Sentiraama" was created by G.Rama Rohit Reddy at Language Technologies Research Centre, KCIS, IIIT Hyderabad. In the corpus, a folder named "Movie Reviews" contains 267 different Telugu movie reviews written in Telugu script. Out of them, 136 are positive and 131 are negative. It contains a total of 20000 sentences and 165049 words. In the corpus, a folder named "Product Reviews" contains 200 different product reviews written in Telugu script. Out of them, 100 are positive and 100 are negative. It contains a total of 43199 sentences and 259189 words, we created our own dataset from this reference of the above statement, the dataset that contains the whole story of a movie so we decided to make a long sentence into a short and we labeled it as "positive" and "negative".
Step-2: Pre-processing Data The collected data is pre-processed using different pre-processing techniques and splitting the large Telugu Sentence into small sentences.
Step-3: Connecting to Hugging Face Hugging Face provides a token with which we can log in using a notebook function and the rest of the work we do will be exported to the platform automatically.
Step-4: Loading pre-trained model and tokenizer The pre-trained model and tokenizer from xlm-roberta-base are loaded for training our Telugu data
Step-5: Training the model Required libraries like Trainer and Training arguments are imported from the Transformers library. After giving the Training arguments with our data we train the model using the train() method which takes 1 to 1 ½ hours depending upon the size of our input data
Step-6: Pushing model and tokenizer Then trainer.push_to_hub() and tokenizer.push_to_hub() methods are used to export our trained model and its tokenizers which are used for the mapping of words in prediction.
Step-7: Testing In the hugging face after opening our model page there is an API in which We give a Telugu Sentence as input with
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
Training results
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
- Downloads last month
- 28