mdeberta-v3-base_ordinal_5_seed420_EN-NL

This model is a fine-tuned version of microsoft/mdeberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.3573
• Mse: 2.7856
• Rmse: 1.6690
• Mae: 0.9730
• R2: 0.2132
• F1: 0.7591
• Precision: 0.7626
• Recall: 0.7648
• Accuracy: 0.7648

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
3.5638	0.2141	100	3.5430	3.6731	1.9165	1.8782	-0.0652	0.4761	0.3859	0.6212	0.6212
3.2282	0.4283	200	3.1915	3.7250	1.9300	1.6743	-0.0802	0.4761	0.3859	0.6212	0.6212
3.0493	0.6424	300	3.0534	3.1846	1.7845	1.6116	0.0765	0.4761	0.3859	0.6212	0.6212
2.9037	0.8565	400	2.9377	3.3112	1.8197	1.4246	0.0398	0.4761	0.3859	0.6212	0.6212
2.7728	1.0707	500	2.7732	3.0639	1.7504	1.3655	0.1115	0.6107	0.6498	0.6586	0.6586
2.6147	1.2848	600	2.6980	3.0434	1.7445	1.3281	0.1174	0.6937	0.6983	0.6912	0.6912
2.5372	1.4989	700	2.5946	3.0024	1.7327	1.2220	0.1293	0.6905	0.6908	0.6984	0.6984
2.4555	1.7131	800	2.4696	2.7563	1.6602	1.1809	0.2007	0.7178	0.7173	0.7226	0.7226
2.4496	1.9272	900	2.4942	2.6828	1.6379	1.2232	0.2220	0.7327	0.7355	0.7310	0.7310
2.298	2.1413	1000	2.5051	3.0072	1.7341	1.1086	0.1279	0.7176	0.7205	0.7262	0.7262
2.2482	2.3555	1100	2.4543	2.8938	1.7011	1.0748	0.1608	0.7215	0.7344	0.7358	0.7358
2.0678	2.5696	1200	2.3826	2.8914	1.7004	1.0338	0.1615	0.7434	0.7460	0.7419	0.7419
2.0865	2.7837	1300	2.3957	2.8504	1.6883	1.0145	0.1734	0.7383	0.7397	0.7443	0.7443
2.1771	2.9979	1400	2.3659	2.7370	1.6544	1.0579	0.2063	0.7438	0.7434	0.7443	0.7443
2.0164	3.2120	1500	2.3783	2.9071	1.7050	1.0398	0.1570	0.7462	0.7497	0.7443	0.7443
1.9577	3.4261	1600	2.4072	2.8902	1.7001	1.0543	0.1619	0.7471	0.7549	0.7443	0.7443
1.8874	3.6403	1700	2.3547	2.7913	1.6707	0.9795	0.1905	0.7523	0.7537	0.7575	0.7575
1.8746	3.8544	1800	2.3244	2.6767	1.6361	1.0024	0.2238	0.7507	0.7499	0.7527	0.7527
1.9356	4.0685	1900	2.3361	2.8166	1.6783	1.0217	0.1832	0.7568	0.7573	0.7563	0.7563
1.7507	4.2827	2000	2.3419	2.7575	1.6606	0.9867	0.2003	0.7506	0.7512	0.7551	0.7551
1.7485	4.4968	2100	2.3295	2.8323	1.6830	1.0109	0.1786	0.7573	0.7587	0.7563	0.7563
1.7336	4.7109	2200	2.3332	2.7262	1.6511	0.9578	0.2094	0.7565	0.7590	0.7624	0.7624
1.739	4.9251	2300	2.4113	2.9928	1.7300	1.0314	0.1321	0.7415	0.7454	0.7394	0.7394
1.6066	5.1392	2400	2.4276	3.0145	1.7362	1.0265	0.1258	0.7613	0.7686	0.7587	0.7587
1.6271	5.3533	2500	2.3737	2.9373	1.7138	0.9783	0.1482	0.7597	0.7595	0.7600	0.7600

Framework versions

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