PDAP/coarse-url-classifier

distilbert_coarse5_js_1.1

This model is a fine-tuned version of distilbert-base-uncased trained on the dataset PDAP/coarse-labeled-urls-headers. It achieves the following results on the evaluation set:

• Loss: 0.6826
• Accuracy: 0.8039

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

This model is trained on urls/html data belonging to 5 coarse grained labels:

• Police & Public Interactions
• Info About Officers
• Info About Agencies
• Agency-Published Resources
• Jails & Courts Specific

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-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
• lr_scheduler_warmup_steps: 100
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	364	0.9021	0.6830
1.0729	2.0	728	0.6936	0.7712
0.6279	3.0	1092	0.6766	0.7745
0.6279	4.0	1456	0.6633	0.7941
0.4531	5.0	1820	0.6691	0.8137
0.3527	6.0	2184	0.6826	0.8039

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2
• Datasets 2.16.1
• Tokenizers 0.15.0