spellcorrector_2410_v14_canine-s

This model is a fine-tuned version of google/canine-s on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0043
• Precision: 0.9996
• Recall: 0.9994
• F1: 0.9995
• Accuracy: 0.9989

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.19	1.0	1951	0.1535	0.9437	0.9787	0.9609	0.9670
0.1486	2.0	3902	0.1187	0.9591	0.9774	0.9682	0.9716
0.126	3.0	5853	0.1029	0.9671	0.9792	0.9731	0.9739
0.1082	4.0	7804	0.0888	0.9740	0.9785	0.9763	0.9769
0.0992	5.0	9755	0.0719	0.9762	0.9852	0.9807	0.9813
0.0871	6.0	11706	0.0624	0.9805	0.9861	0.9833	0.9832
0.0782	7.0	13657	0.0527	0.9835	0.9885	0.9860	0.9858
0.0693	8.0	15608	0.0446	0.9866	0.9898	0.9882	0.9876
0.0604	9.0	17559	0.0375	0.9888	0.9906	0.9897	0.9893
0.0543	10.0	19510	0.0318	0.9915	0.9926	0.9921	0.9914
0.046	11.0	21461	0.0272	0.9932	0.9940	0.9936	0.9925
0.0425	12.0	23412	0.0217	0.9942	0.9950	0.9946	0.9939
0.0378	13.0	25363	0.0188	0.9953	0.9963	0.9958	0.9946
0.0333	14.0	27314	0.0160	0.9963	0.9962	0.9962	0.9954
0.0286	15.0	29265	0.0140	0.9972	0.9970	0.9971	0.9960
0.0261	16.0	31216	0.0121	0.9977	0.9978	0.9978	0.9966
0.0235	17.0	33167	0.0104	0.9984	0.9979	0.9982	0.9972
0.021	18.0	35118	0.0090	0.9987	0.9986	0.9986	0.9976
0.0196	19.0	37069	0.0073	0.9990	0.9988	0.9989	0.9980
0.0166	20.0	39020	0.0064	0.9992	0.9991	0.9991	0.9983
0.0158	21.0	40971	0.0059	0.9994	0.9991	0.9992	0.9984
0.0136	22.0	42922	0.0053	0.9995	0.9994	0.9994	0.9986
0.0134	23.0	44873	0.0047	0.9996	0.9993	0.9994	0.9988
0.0125	24.0	46824	0.0045	0.9996	0.9993	0.9995	0.9989
0.0116	25.0	48775	0.0043	0.9996	0.9994	0.9995	0.9989

Framework versions

• Transformers 4.28.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.13.3