mnavas / roberta-finetuned-WebClassification

roberta-finetuned-WebClassification

This model is a fine-tuned version of xlm-roberta-base on the Web Classification Dataset . It achieves the following results on the evaluation set:

• Loss: 0.3473
• Accuracy: 0.9504
• F1: 0.9504
• Precision: 0.9504
• Recall: 0.9504

Model description

The model classifies websites into the following categories:

• "0": "Adult",
• "1": "Business/Corporate",
• "2": "Computers and Technology",
• "3": "E-Commerce",
• "4": "Education",
• "5": "Food",
• "6": "Forums",
• "7": "Games",
• "8": "Health and Fitness",
• "9": "Law and Government",
• "10": "News",
• "11": "Photography",
• "12": "Social Networking and Messaging",
• "13": "Sports",
• "14": "Streaming Services",
• "15": "Travel"

Intended uses & limitations

Web classification in English (for now).

Training and evaluation data

Trained and tested on a 80/20 split of the Web Classification Dataset .

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Framework versions

• Transformers 4.16.2
• Pytorch 1.9.1
• Datasets 1.18.4
• Tokenizers 0.11.6