fine-tunning-roberta-bne-hate-offensive

This model is a fine-tuned version of BSC-LT/roberta-base-bne on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6211
• Hate/offensive F1: 0.8958
• Hate/offensive Precision: 0.9421
• Hate/offensive Recall: 0.8539
• Normal F1: 0.9519
• Normal Precision: 0.9308
• Normal Recall: 0.9740
• Micro F1: 0.9342
• Acc: 0.9342
• Macro F1: 0.9239
• Macro Precision: 0.9364
• Macro Recall: 0.9139

Model description

More information needed

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: 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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Hate/offensive F1	Hate/offensive Precision	Hate/offensive Recall	Normal F1	Normal Precision	Normal Recall	Micro F1	Acc	Macro F1	Macro Precision	Macro Recall
0.2865	1.0	1064	0.3398	0.8953	0.9324	0.8610	0.9510	0.9336	0.9691	0.9332	0.9332	0.9231	0.9330	0.9150
0.1527	2.0	2128	0.3417	0.8942	0.9204	0.8695	0.9497	0.9370	0.9627	0.9318	0.9318	0.9220	0.9287	0.9161
0.047	3.0	3192	0.5626	0.8914	0.9668	0.8270	0.9518	0.9199	0.9859	0.9332	0.9332	0.9216	0.9434	0.9064
0.0145	4.0	4256	0.5874	0.8986	0.9396	0.8610	0.9528	0.9338	0.9726	0.9356	0.9356	0.9257	0.9367	0.9168
0.0029	5.0	5320	0.6211	0.8958	0.9421	0.8539	0.9519	0.9308	0.9740	0.9342	0.9342	0.9239	0.9364	0.9139

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0