roberta-issue-classifier

This model is a fine-tuned version of roberta-base on this dataset. It achieves the following results on the evaluation set:

• Loss: 0.0945
• Accuracy: 0.9844
• F1: 0.9844

Model description

Issue Type Mapping: {'economic': 0, 'environmental': 1, 'family/gender': 2, 'geo-political and foreign policy': 3, 'political': 4, 'racial justice and immigration': 5, 'religious': 6, 'social, health and education': 7}

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.5661	0.625	100	0.4350	0.9437	0.9436
0.112	1.25	200	0.1488	0.975	0.9750
0.0335	1.875	300	0.1262	0.9781	0.9781
0.1009	2.5	400	0.1328	0.9781	0.9781
0.032	3.125	500	0.0945	0.9844	0.9844
0.0074	3.75	600	0.0944	0.9781	0.9781

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0