File size: 2,306 Bytes
3617524 d6e6809 3617524 d6e6809 3617524 | 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 | ---
library_name: transformers
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: toxicity-classification-model
results: []
datasets:
- dirtycomputer/Toxic_Comment_Classification_Challenge
language:
- en
pipeline_tag: text-classification
---
# toxicity-classification-model
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the dirtycomputer/Toxic_Comment_Classification_Challenge dataset. It achieves the following results on the evaluation set:
- Loss: 0.0511
- Accuracy: 0.9812
## Model description
Fine-tuned roberta-base model for detecting toxicity in comments. It categorizes a comment into different toxicity types, such as "toxic," "obscene," "insult," and "threat."
## Intended uses & limitations
### Intended Uses
- **Content Moderation**: Automatically flagging or removing toxic comments on social media platforms, forums, and customer support.
- **Toxicity Detection**: Classifying comments based on their toxicity level, such as harmful language or insults.
### Limitations
- **False Negatives**: May not always catch subtle toxic behavior.
- **Limited Language Support**: Currently, the model is trained on English-only data.
- **Context Sensitivity**: May struggle with ambiguous language or sarcasm.
## Training and evaluation data
This model was fine-tuned using the **dirtycomputer/Toxic_Comment_Classification_Challenge** dataset, which contains labeled comments for toxicity classification.
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 16
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9, 0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.1691 | 1.0 | 17952 | 0.1464 | 0.9617 |
| 0.0892 | 2.0 | 35904 | 0.1456 | 0.9617 |
| 0.0527 | 3.0 | 53856 | 0.0511 | 0.9812 |
### Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0 |