classification-hate-speech-DE-7epoch-idblp2

This model is a fine-tuned version of indolem/indobertweet-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4500
• F1 macro: 0.3864
• Weighted: 0.5746
• Balanced accuracy: 0.5221

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 14

Training results

Training Loss	Epoch	Step	Validation Loss	F1 macro	Weighted	Balanced accuracy
1.5725	1.0	38	1.3996	0.3389	0.4577	0.4681
1.0621	2.0	76	1.3070	0.3838	0.5860	0.4956
0.6859	3.0	114	1.2224	0.3973	0.6523	0.5086
0.3201	4.0	152	1.6724	0.3774	0.5827	0.5124
0.174	5.0	190	1.5607	0.4048	0.6200	0.5299
0.0721	6.0	228	1.6479	0.3935	0.6215	0.5118
0.0459	7.0	266	2.3813	0.3669	0.5297	0.5219
0.0164	8.0	304	2.2315	0.3857	0.5657	0.5242
0.0062	9.0	342	2.1938	0.3899	0.5802	0.5288
0.0074	10.0	380	2.4419	0.3825	0.5629	0.5217
0.004	11.0	418	2.3171	0.3897	0.5812	0.5233
0.0031	12.0	456	2.3685	0.3873	0.5760	0.5223
0.0032	13.0	494	2.4116	0.3890	0.5753	0.5238
0.0039	14.0	532	2.4500	0.3864	0.5746	0.5221

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1