945323ddea7099fba8e371750bdf1698

This model is a fine-tuned version of albert/albert-xlarge-v1 on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.3181
• Data Size: 1.0
• Epoch Runtime: 40.9352
• Accuracy: 0.9030
• F1 Macro: 0.6648

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	1.5254	0	3.4868	0.2078	0.1338
No log	1	619	0.6702	0.0078	3.8240	0.7672	0.2894
No log	2	1238	0.6159	0.0156	3.8973	0.7675	0.2902
0.0162	3	1857	0.5463	0.0312	4.6526	0.7823	0.5004
0.0162	4	2476	0.5718	0.0625	5.7954	0.8228	0.4914
0.3715	5	3095	0.3547	0.125	7.8975	0.8868	0.6633
0.0304	6	3714	0.3074	0.25	12.5095	0.8963	0.5959
0.3091	7	4333	0.2964	0.5	22.1105	0.8983	0.6669
0.2872	8.0	4952	0.2738	1.0	41.7306	0.9040	0.6060
0.2406	9.0	5571	0.2973	1.0	40.7432	0.8981	0.7062
0.2419	10.0	6190	0.2710	1.0	41.2123	0.9040	0.7011
0.2524	11.0	6809	0.2965	1.0	40.9208	0.9034	0.7298
0.2118	12.0	7428	0.2955	1.0	41.5401	0.9073	0.6532
0.2259	13.0	8047	0.2784	1.0	40.9850	0.9079	0.7421
0.1876	14.0	8666	0.3181	1.0	40.9352	0.9030	0.6648

Framework versions

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