nb-sbert-base-edu-scorer-lr3e4-bs32-swe

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

• Loss: 0.7479
• Mse: 0.7479
• Mae: 0.6630
• Rmse: 0.8648
• R2: 0.6113

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: 0.0003
• train_batch_size: 32
• eval_batch_size: 32
• 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: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mse	Mae	Rmse	R2
No log	0	0	6.8805	6.8805	2.1967	2.6231	-2.4611
0.9599	0.3397	1000	0.9498	0.9498	0.7640	0.9746	0.5222
0.9341	0.6793	2000	0.9336	0.9336	0.7346	0.9662	0.5303
0.9028	1.0190	3000	0.8837	0.8837	0.7160	0.9400	0.5555
0.9487	1.3587	4000	0.9146	0.9146	0.7252	0.9563	0.5399
0.9194	1.6984	5000	0.8668	0.8668	0.7192	0.9310	0.5640
0.8753	2.0380	6000	0.8623	0.8623	0.7353	0.9286	0.5662
0.8543	2.3777	7000	0.8650	0.8650	0.7066	0.9300	0.5649
0.8819	2.7174	8000	0.8821	0.8821	0.7162	0.9392	0.5563
0.8642	3.0571	9000	0.8859	0.8859	0.7103	0.9412	0.5544
0.8439	3.3967	10000	0.8715	0.8715	0.7073	0.9335	0.5616
0.8415	3.7364	11000	0.8568	0.8568	0.7036	0.9256	0.5690
0.8587	4.0761	12000	0.8363	0.8363	0.7023	0.9145	0.5793
0.8756	4.4158	13000	0.9141	0.9141	0.7212	0.9561	0.5402
0.821	4.7554	14000	0.8662	0.8662	0.7027	0.9307	0.5643
0.7879	5.0951	15000	0.8590	0.8590	0.7267	0.9268	0.5679
0.8004	5.4348	16000	0.8575	0.8575	0.7049	0.9260	0.5687
0.8436	5.7745	17000	0.8379	0.8379	0.7079	0.9154	0.5785
0.8116	6.1141	18000	0.8296	0.8296	0.7003	0.9108	0.5827
0.8027	6.4538	19000	0.8433	0.8433	0.6905	0.9183	0.5758
0.8269	6.7935	20000	0.8274	0.8274	0.6951	0.9096	0.5838
0.8064	7.1332	21000	0.8359	0.8359	0.7032	0.9143	0.5795
0.8019	7.4728	22000	0.8230	0.8230	0.6902	0.9072	0.5860
0.8072	7.8125	23000	0.8227	0.8227	0.7104	0.9071	0.5861
0.8097	8.1522	24000	0.8859	0.8859	0.7086	0.9412	0.5544
0.785	8.4918	25000	0.8272	0.8272	0.6989	0.9095	0.5839
0.796	8.8315	26000	0.8166	0.8166	0.6867	0.9037	0.5892
0.8285	9.1712	27000	0.8280	0.8280	0.6945	0.9100	0.5835
0.7995	9.5109	28000	0.8351	0.8351	0.6897	0.9138	0.5799
0.8177	9.8505	29000	0.8191	0.8191	0.6976	0.9051	0.5880
0.7801	10.1902	30000	0.8114	0.8114	0.6940	0.9008	0.5918
0.7986	10.5299	31000	0.8139	0.8139	0.6885	0.9022	0.5906
0.7764	10.8696	32000	0.8062	0.8062	0.6950	0.8979	0.5944
0.7747	11.2092	33000	0.8276	0.8276	0.6859	0.9097	0.5837
0.7761	11.5489	34000	0.8260	0.8260	0.6943	0.9088	0.5845
0.7436	11.8886	35000	0.8186	0.8186	0.6873	0.9048	0.5882
0.7461	12.2283	36000	0.8164	0.8164	0.6881	0.9035	0.5893
0.8033	12.5679	37000	0.8160	0.8160	0.6867	0.9033	0.5895
0.7739	12.9076	38000	0.8077	0.8077	0.6875	0.8987	0.5937
0.7596	13.2473	39000	0.8180	0.8180	0.6882	0.9045	0.5885
0.7938	13.5870	40000	0.8066	0.8066	0.6921	0.8981	0.5942
0.7822	13.9266	41000	0.8264	0.8264	0.6891	0.9091	0.5843
0.7615	14.2663	42000	0.8117	0.8117	0.6859	0.9009	0.5917
0.7912	14.6060	43000	0.8358	0.8358	0.6894	0.9142	0.5796
0.7675	14.9457	44000	0.8020	0.8020	0.6843	0.8955	0.5966
0.7117	15.2853	45000	0.8182	0.8182	0.6892	0.9045	0.5884
0.7439	15.625	46000	0.8023	0.8023	0.6912	0.8957	0.5964
0.7614	15.9647	47000	0.8020	0.8020	0.6898	0.8955	0.5966
0.7591	16.3043	48000	0.8145	0.8145	0.6896	0.9025	0.5903
0.7949	16.6440	49000	0.8129	0.8129	0.6845	0.9016	0.5911
0.747	16.9837	50000	0.8093	0.8093	0.6840	0.8996	0.5929
0.752	17.3234	51000	0.8071	0.8071	0.6898	0.8984	0.5940
0.7236	17.6630	52000	0.8027	0.8027	0.6830	0.8959	0.5962
0.7392	18.0027	53000	0.8081	0.8081	0.6831	0.8990	0.5935
0.7154	18.3424	54000	0.8057	0.8057	0.6867	0.8976	0.5947
0.766	18.6821	55000	0.8104	0.8104	0.6837	0.9002	0.5923
0.7268	19.0217	56000	0.8061	0.8061	0.6849	0.8978	0.5945
0.7886	19.3614	57000	0.8074	0.8074	0.6859	0.8986	0.5939
0.7258	19.7011	58000	0.8051	0.8051	0.6856	0.8973	0.5950

Framework versions

• Transformers 4.55.0
• Pytorch 2.5.1+cu121
• Datasets 4.0.0
• Tokenizers 0.21.4