hate_speech_detection_with_target-bert-large-portuguese-cased

This model is a fine-tuned version of neuralmind/bert-large-portuguese-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0610
• Accuracy: 0.9892

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.8021	1.0	93	0.1410	0.9594
0.0974	2.0	186	0.0974	0.9770
0.1401	3.0	279	0.0550	0.9838
0.046	4.0	372	0.0618	0.9878
0.0344	5.0	465	0.0469	0.9892
0.2429	6.0	558	0.0854	0.9878
0.0696	7.0	651	0.0451	0.9892
0.0394	8.0	744	0.0460	0.9892
0.0279	9.0	837	0.0469	0.9892
0.0362	10.0	930	0.0779	0.9865
0.0215	11.0	1023	0.0655	0.9878
0.0193	12.0	1116	0.0587	0.9892
0.0154	13.0	1209	0.0594	0.9892
0.015	14.0	1302	0.0601	0.9905
0.0156	15.0	1395	0.0604	0.9892
0.0157	16.0	1488	0.0604	0.9892
0.0145	17.0	1581	0.0607	0.9892
0.0176	18.0	1674	0.0607	0.9892
0.0193	19.0	1767	0.0609	0.9892
0.0206	20.0	1860	0.0610	0.9892

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.0+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0