768df56b99429d0add3fb177c05296ff

This model is a fine-tuned version of distilbert/distilbert-base-multilingual-cased on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.4378
• Data Size: 1.0
• Epoch Runtime: 20.3831
• Accuracy: 0.9091
• F1 Macro: 0.7151

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.1552	0	1.9155	0.0629	0.0562
No log	1	619	0.7090	0.0078	2.4594	0.7672	0.2894
No log	2	1238	0.6622	0.0156	2.6005	0.7672	0.2894
0.0158	3	1857	0.4367	0.0312	2.9326	0.8506	0.5331
0.0158	4	2476	0.3438	0.0625	3.4452	0.8797	0.5792
0.3265	5	3095	0.3101	0.125	4.6903	0.8904	0.6569
0.0249	6	3714	0.2976	0.25	7.2107	0.9012	0.6273
0.2678	7	4333	0.2658	0.5	12.1906	0.9014	0.7480
0.2351	8.0	4952	0.2756	1.0	22.4078	0.9093	0.6782
0.168	9.0	5571	0.2892	1.0	21.4293	0.9067	0.7640
0.1469	10.0	6190	0.3359	1.0	20.9939	0.9095	0.7436
0.0978	11.0	6809	0.4378	1.0	20.3831	0.9091	0.7151

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1