File size: 1,826 Bytes
9adcadd |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 |
---
license: afl-3.0
language:
- en
- te
metrics:
- accuracy
pipeline_tag: text-classification
library_name: transformers
tags:
- toxic-comment-classification
- roberta
- text-classification
---
# Toxic Comment Classification Using RoBERTa
## Overview
This project provides a toxic comment classification model based on RoBERTa (Robustly optimized BERT approach). The model is designed to classify comments as toxic or non-toxic, helping in moderating online discussions and improving community interactions.
## Model Details
- **Model Name**: RoBERTa for Toxic Comment Classification
- **Architecture**: RoBERTa
- **Fine-tuning Task**: Binary classification (toxic vs. non-toxic)
- **Evaluation Metrics**:
- Accuracy
- F1 Score
- Precision
- Recall
## Files
- `pytorch_model.bin`: The trained model weights.
- `config.json`: Model configuration file.
- `merges.txt`: BPE tokenizer merges file.
- `model.safetensors`: Model weights in safetensors format.
- `special_tokens_map.json`: Tokenizer special tokens mapping.
- `tokenizer_config.json`: Tokenizer configuration file.
- `vocab.json`: Tokenizer vocabulary file.
- `roberta-toxic-comment-classifier.pkl`: Serialized best model state dictionary (for PyTorch).
- `README.md`: This documentation file.
## Model Performance
- **Accuracy**: 0.9599
- **F1 Score**: 0.9615
- **Precision**: 0.9646
- **Recall**: 0.9599
## Load the model
```
from transformers import pipeline
# Load the model and tokenizer
model_name = "prabhaskenche/pk-toxic-comment-classification-using-RoBERTa"
classifier = pipeline("text-classification", model=model_name)
# Example usage
text = "You're the worst person I've ever met."
result = classifier(text)
print(result)
```
## Usage
### Installation
Install the required packages:
```bash
pip install torch transformers sklearn |