EuroBERT-210m-humour-detection

This model is a fine-tuned version of EuroBERT/EuroBERT-210m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4992
• Accuracy: 0.7665
• Precision: 0.7514
• Recall: 0.8029
• F1: 0.7763

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: 16
• eval_batch_size: 16
• 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: cosine
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.3213	0.2222	500	0.2204	0.9176	0.9153	0.9205	0.9179
0.2453	0.4444	1000	0.2832	0.9202	0.8754	0.9799	0.9247
0.2058	0.6667	1500	0.1405	0.9389	0.9626	0.9133	0.9373
0.1893	0.8889	2000	0.1722	0.9407	0.9156	0.9709	0.9424
0.1533	1.1111	2500	0.1565	0.9483	0.9390	0.9588	0.9488
0.1324	1.3333	3000	0.1516	0.9474	0.9343	0.9624	0.9482
0.1283	1.5556	3500	0.1512	0.9494	0.9612	0.9367	0.9488
0.1209	1.7778	4000	0.1489	0.9458	0.9634	0.9269	0.9448
0.1196	2.0	4500	0.1538	0.9505	0.9467	0.9547	0.9507
0.0884	2.2222	5000	0.2242	0.9470	0.9605	0.9323	0.9462
0.0822	2.4444	5500	0.2198	0.9472	0.9591	0.9344	0.9466
0.0871	2.6667	6000	0.2110	0.9471	0.9538	0.9398	0.9467
0.0726	2.8889	6500	0.2161	0.9480	0.9550	0.9403	0.9476

Framework versions

• Transformers 4.52.2
• Pytorch 2.8.0.dev20250521
• Datasets 3.6.0
• Tokenizers 0.21.1