AbhiSajeev/sentiment-analysis-t4

Sentiment Analysis Model

Model Description

This model is a fine-tuned version of DistilBERT for binary sentiment classification. The model was trained on a custom sentiment dataset containing 100 labeled text samples.

Model Details

Developed by: Abhinava Sajeev Model Type: DistilBERT Sequence Classification Language: English Base Model: distilbert-base-uncased Task: Sentiment Classification Labels:

0 = Negative 1 = Positive

Intended Use

This model predicts whether a given English text expresses positive or negative sentiment.

Example

Input:

"This product is amazing!"

Output:

Positive

Training Data

The model was trained on a custom dataset consisting of:

100 text samples 50 positive samples 50 negative samples

Dataset columns:

text – Input text label – Sentiment label

Training Configuration

Epochs: 3 Batch Size: 8 Max Sequence Length: 128 Optimizer: AdamW Framework: Hugging Face Transformers Hardware: NVIDIA Tesla T4 GPU Platform: Google Colab

Evaluation

The dataset was split into:

80% Training Data 20% Testing Data

The model achieved satisfactory performance for a small educational dataset.

Usage

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model="AbhiSajeev/sentiment-analysis-t4"
)

result = classifier("I really enjoyed using this application.")
print(result)

Limitations

The model was trained on only 100 examples. It is intended for educational and demonstration purposes. Performance may not generalize well to unseen domains.

Technical Specifications

Architecture

DistilBERT Transformer Encoder Binary Classification

Software

Python PyTorch Transformers Datasets Accelerate

Author

Abhinava Sajeev

Contact

Abhinava Sajeev