bertweet-base_ordinal_5_seed42_EN

This model is a fine-tuned version of vinai/bertweet-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.3762
• Mse: 2.9969
• Rmse: 1.7311
• Mae: 0.9259
• R2: 0.1490
• F1: 0.7832
• Precision: 0.7840
• Recall: 0.7860
• Accuracy: 0.7860

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-06
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Mse	Rmse	Mae	R2	F1	Precision	Recall	Accuracy
3.5232	0.4630	100	3.4011	3.4987	1.8705	1.7180	-0.0102	0.4570	0.3669	0.6057	0.6057
3.2579	0.9259	200	3.1630	3.3081	1.8188	1.6110	0.0448	0.4570	0.3669	0.6057	0.6057
2.9868	1.3889	300	2.9383	3.4125	1.8473	1.3916	0.0146	0.4570	0.3669	0.6057	0.6057
2.7421	1.8519	400	2.7612	3.0757	1.7538	1.2167	0.1119	0.6594	0.7657	0.7076	0.7076
2.4831	2.3148	500	2.5697	2.7520	1.6589	1.1070	0.2054	0.7538	0.7580	0.7598	0.7598
2.3685	2.7778	600	2.5183	2.9478	1.7169	1.0888	0.1488	0.7405	0.7549	0.7520	0.7520
2.2204	3.2407	700	2.4655	2.9530	1.7184	0.9739	0.1473	0.7501	0.7776	0.7650	0.7650
2.0367	3.7037	800	2.4423	3.1462	1.7738	0.9843	0.0915	0.7546	0.7546	0.7546	0.7546
2.0266	4.1667	900	2.3620	2.9452	1.7161	0.9608	0.1496	0.7656	0.7653	0.7676	0.7676
1.8431	4.6296	1000	2.3728	3.0157	1.7366	0.9373	0.1292	0.7680	0.7679	0.7702	0.7702
1.8448	5.0926	1100	2.3544	2.8903	1.7001	0.9164	0.1654	0.7754	0.7759	0.7781	0.7781
1.7661	5.5556	1200	2.4342	3.0026	1.7328	0.9138	0.1330	0.7799	0.7794	0.7807	0.7807
1.6105	6.0185	1300	2.4327	3.1227	1.7671	0.9347	0.0983	0.7691	0.7686	0.7702	0.7702
1.5606	6.4815	1400	2.4423	3.0966	1.7597	0.9347	0.1059	0.7732	0.7737	0.7728	0.7728

Framework versions

• Transformers 4.40.2
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.19.1